------------------------------------------------------------------------------------------------------------------
      name:  <unnamed>
       log:  C:\Users\mbeissin\Desktop\Stata files for book\Robustnesstestfiles\Logfiles\robustnesstestschapter3.l
> og
  log type:  text
 opened on:  26 Jan 2022, 09:19:54

. * ============================================================================
. * ROBUSTNESS CHECKS FOR STATISTICAL RESULTS APPEARING IN CHAPTER 3
. * STATA Do file for Chapter 3 
. * Robustness checks for results reported in Chapter 3  
. * Author: Mark R. Beissinger  
. * Date:  January 2022  
. * Princeton, NJ 
. * =============================================================================
. * BEFORE RUNNING, YOU MUST SET THE DEFAULT PATH FOR WHERE THE DATA
. *   FILES RESIDE
. * ============================================================================
. * The following datafile is used in this file:
. *   Panel data for revolutionary episodes--revspredictbycntryyr.dta
. * ============================================================================
. * Before running, download the following packages for STATA:
. * firthlogit from http://fmwww.bc.edu/RePEc/bocode/f
. * relogit from https://gking.harvard.edu/relogit
. * checkrob from http://fmwww.bc.edu/RePEc/bocode/c
. * qic (as discussed in https://www.stata-journal.com/article.html?article=st0126)
. *       --install from http://www.stata-journal.com/software/sj8-1/
. * =============================================================================
. * The following output is produced by these robustness tests:
. *               Robustnesstestfiles\Logfiles\robustnesstestschapter3.log
. *
. *               In addition, the following graphs of imputed vs. observed observa-
. *                       tions were produced:
. *                               Robustnesstestfiles\Logfiles\mod4impobsunder5mortl.pdf
. *                               Robustnesstestfiles\Logfiles\mod4impobspercurbanl.pdf
. *                               Robustnesstestfiles\Logfiles\mod4impobstotalyrsschooll.pdf
. *                               Robustnesstestfiles\Logfiles\mod4impobsmilperthousl.pdf
. *                               Robustnesstestfiles\Logfiles\mod4impobsyouthpercl.pdf
. *                               Robustnesstestfiles\Logfiles\mod4impobstottradepernomgdpl.pdf
. *                               Robustnesstestfiles\Logfiles\mod4impobslndollexchratel.pdf
. *                               Robustnesstestfiles\Logfiles\mod4impobsageincleader.pdf
. *                               Robustnesstestfiles\Logfiles\mod4impobslnmilexppersoldthl.pdf
. *       These files have been combined into a single pdf version of the
. *                       output file, located in the Robustnesstestfiles\Outputfiles folder
. *       In addition, the reworked output from the checkrob procedure run in this
. *          chapter can be viewed in the Excel file checkrob.results.chapter3.xlsx,
. *               also located in the Robustnesstestfiles\Outputfiles folder      
. * =============================================================================
. 
. use revspredictbycntryyr.dta

. 
. * =====================================================================================
. * ROBUSTNESS CHECKS: CHECKING QUADRATURE FOR BIVARIATE REGRESSIONS (FIGURES 3.1 TO 3.4)
. * =====================================================================================
. * Figure 3.1. Polity score
. * Urban civic
. xtcloglog urbancivicny  polityl c.polityl#c.polityl c.time1##c.time1##c.time1 if indstate==1, vce(robust) eform 
> nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,328
Group variable: cowcode                         Number of groups  =        159

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
                                                              avg =       71.2
                                                              max =        115

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(5)      =      45.58
Log pseudolikelihood  = -286.70449              Prob > chi2       =     0.0000

                                         (Std. Err. adjusted for 159 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
           urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
                polityl |   .9097204   .0249435    -3.45   0.001     .8621224    .9599464
                        |
    c.polityl#c.polityl |   .9836101   .0049711    -3.27   0.001     .9739151    .9934016
                        |
                  time1 |   .9544682   .0784993    -0.57   0.571     .8123725    1.121418
                        |
        c.time1#c.time1 |   1.001478    .001393     1.06   0.288     .9987515    1.004212
                        |
c.time1#c.time1#c.time1 |   .9999932   6.95e-06    -0.98   0.325     .9999795    1.000007
                        |
                  _cons |   .0005532   .0009291    -4.47   0.000     .0000206    .0148737
------------------------+----------------------------------------------------------------
               /lnsig2u |  -.8144803   .8784393                      -2.53619    .9072291
------------------------+----------------------------------------------------------------
                sigma_u |   .6654844   .2922938                      .2813672    1.573991
                    rho |   .2121222   .1468104                      .0459181    .6009749
-----------------------------------------------------------------------------------------

. * Quadrature test
. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -286.70449     -286.70447     -286.70449
likelihood                   .00002103     -3.912e-07   Difference
                            -7.336e-08      1.364e-09   Relative difference
-----------------------------------------------------
urbancivicny:-.09461795     -.09461795     -.09461792
 polityl                     4.669e-12      2.593e-08   Difference
                            -4.935e-11     -2.741e-07   Relative difference
-----------------------------------------------------
urbancivicny:-.01652568     -.01652568     -.01652568
c.polityl#~l                 3.238e-13      1.242e-09   Difference
                            -1.959e-11     -7.518e-08   Relative difference
-----------------------------------------------------
urbancivicny:-.04660099     -.04660099     -.04660097
   time1                     4.433e-12      2.412e-08   Difference
                            -9.513e-11     -5.176e-07   Relative difference
-----------------------------------------------------
urbancivicny: .00147696      .00147696      .00147696
c.time1#c.~1                -8.007e-14     -3.925e-10   Difference
                            -5.421e-11     -2.657e-07   Relative difference
-----------------------------------------------------
urbancivicny:-6.839e-06     -6.839e-06     -6.839e-06
c.time1#c.~1                 3.526e-16      1.566e-12   Difference
                            -5.156e-11     -2.290e-07   Relative difference
-----------------------------------------------------
urbancivicny: -7.499734      -7.499734     -7.4997334
   _cons                     9.024e-11      6.142e-07   Difference
                            -1.203e-11     -8.189e-08   Relative difference
-----------------------------------------------------
lnsig2u:     -.81448027     -.81448027     -.81448463
   _cons                    -5.261e-10     -4.362e-06   Difference
                             6.459e-10      5.355e-06   Relative difference
-----------------------------------------------------

. * --Passed:  all coefficients change by less than .01 
. * Social
. xtcloglog leftistny polityl c.time1##c.time1##c.time1 if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,328
Group variable: cowcode                         Number of groups  =        159

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
                                                              avg =       71.2
                                                              max =        115

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      32.63
Log pseudolikelihood  = -368.40781              Prob > chi2       =     0.0000

                                         (Std. Err. adjusted for 159 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
              leftistny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
                polityl |   1.010763   .0206784     0.52   0.601     .9710356    1.052115
                  time1 |   .9587571   .0313343    -1.29   0.198     .8992686    1.022181
                        |
        c.time1#c.time1 |   1.001584   .0007652     2.07   0.038     1.000085    1.003085
                        |
c.time1#c.time1#c.time1 |    .999986   4.92e-06    -2.85   0.004     .9999763    .9999956
                        |
                  _cons |    .007374   .0033865   -10.69   0.000     .0029977    .0181392
------------------------+----------------------------------------------------------------
               /lnsig2u |  -.7080202   .6085355                     -1.900728    .4846875
------------------------+----------------------------------------------------------------
                sigma_u |   .7018679   .2135558                      .3866003    1.274232
                    rho |   .2304591   .1079224                      .0832926    .4967468
-----------------------------------------------------------------------------------------

. * Quadrature test
. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -368.40781     -368.40766     -368.40781
likelihood                   .00015759      1.646e-07   Difference
                            -4.278e-07     -4.469e-10   Relative difference
-----------------------------------------------------
leftistny:    .01070522      .01070522       .0107052
 polityl                     7.083e-15     -1.830e-08   Difference
                             6.616e-13     -1.709e-06   Relative difference
-----------------------------------------------------
leftistny:    -.0421175      -.0421175      -.0421175
   time1                     4.075e-14      9.783e-10   Difference
                            -9.674e-13     -2.323e-08   Relative difference
-----------------------------------------------------
leftistny:    .00158261      .00158261      .00158261
c.time1#c.~1                -7.839e-16      1.763e-11   Difference
                            -4.953e-13      1.114e-08   Relative difference
-----------------------------------------------------
leftistny:   -.00001402     -.00001402     -.00001402
c.time1#c.~1                 2.692e-18     -5.465e-13   Difference
                            -1.921e-13      3.899e-08   Relative difference
-----------------------------------------------------
leftistny:    -4.909801      -4.909801     -4.9097983
   _cons                     7.104e-12      2.677e-06   Difference
                            -1.447e-12     -5.453e-07   Relative difference
-----------------------------------------------------
lnsig2u:     -.70802021     -.70802021     -.70802962
   _cons                    -2.019e-11     -9.409e-06   Difference
                             2.852e-11      .00001329   Relative difference
-----------------------------------------------------

. * --Passed:  all coefficients change by less than .01 
. 
. * On relationship of non-democratic regime-types to probability of onset for urban civic and social revolutionar
> y episodes (Geddes data)
. * Urban civic
. xtcloglog urbancivicny  gedpartyautoc gedmilautoc gedmonautoc gedpersautoc c.time1##c.time1##c.time1 if indstate
> ==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =      7,699
Group variable: cowcode                         Number of groups  =        148

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         17
                                                              avg =       52.0
                                                              max =         65

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(7)      =     101.45
Log pseudolikelihood  = -226.49849              Prob > chi2       =     0.0000

                                         (Std. Err. adjusted for 148 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
           urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
          gedpartyautoc |   4.185325   2.137913     2.80   0.005     1.537891    11.39024
            gedmilautoc |   16.82953   8.762122     5.42   0.000     6.066006    46.69188
            gedmonautoc |   3.093933   3.643171     0.96   0.337     .3077538    31.10416
           gedpersautoc |   6.316402    2.95918     3.93   0.000     2.521675     15.8216
                  time1 |   .5916528   .3517203    -0.88   0.377     .1845238    1.897061
                        |
        c.time1#c.time1 |   1.007509   .0076581     0.98   0.325     .9926108    1.022631
                        |
c.time1#c.time1#c.time1 |   .9999689   .0000316    -0.98   0.325     .9999069    1.000031
                        |
                  _cons |   16.45375   248.4718     0.19   0.853     2.30e-12    1.18e+14
------------------------+----------------------------------------------------------------
               /lnsig2u |  -1.057451     1.3848                      -3.77161    1.656708
------------------------+----------------------------------------------------------------
                sigma_u |   .5893556   .4080699                      .1517069    2.289547
                    rho |   .1743435    .199339                      .0137984    .7611523
-----------------------------------------------------------------------------------------

. * Quadrature test
. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -226.49849     -226.49849     -226.49849
likelihood                   1.739e-06     -1.439e-08   Difference
                            -7.679e-09      6.353e-11   Relative difference
-----------------------------------------------------
urbancivicny: 1.4315845      1.4315845      1.4315844
gedpartyau~c                -3.649e-10     -3.490e-08   Difference
                            -2.549e-10     -2.438e-08   Relative difference
-----------------------------------------------------
urbancivicny: 2.8231353      2.8231353      2.8231353
gedmilautoc                 -4.001e-10     -1.723e-08   Difference
                            -1.417e-10     -6.102e-09   Relative difference
-----------------------------------------------------
urbancivicny: 1.1294432      1.1294432      1.1294432
gedmonautoc                 -2.940e-10     -2.778e-08   Difference
                            -2.603e-10     -2.459e-08   Relative difference
-----------------------------------------------------
urbancivicny: 1.8431497      1.8431497      1.8431497
gedpersautoc                -2.570e-10     -3.005e-08   Difference
                            -1.394e-10     -1.630e-08   Relative difference
-----------------------------------------------------
urbancivicny:-.52483537     -.52483537     -.52483536
   time1                     2.252e-10      1.704e-08   Difference
                            -4.290e-10     -3.246e-08   Relative difference
-----------------------------------------------------
urbancivicny: .00748101      .00748101      .00748101
c.time1#c.~1                -2.897e-12     -2.289e-10   Difference
                            -3.872e-10     -3.059e-08   Relative difference
-----------------------------------------------------
urbancivicny:-.00003113     -.00003113     -.00003113
c.time1#c.~1                 1.163e-14      9.542e-13   Difference
                            -3.737e-10     -3.065e-08   Relative difference
-----------------------------------------------------
urbancivicny: 2.8005533      2.8005533       2.800553
   _cons                    -4.482e-09     -3.008e-07   Difference
                            -1.600e-09     -1.074e-07   Relative difference
-----------------------------------------------------
lnsig2u:      -1.057451      -1.057451     -1.0574514
   _cons                    -3.618e-09     -4.481e-07   Difference
                             3.421e-09      4.237e-07   Relative difference
-----------------------------------------------------

. * --Passed:  all coefficients change by less than .01 
. * Social
. xtcloglog leftistny  gedpartyautoc gedmilautoc gedmonautoc gedpersautoc c.time1##c.time1##c.time1 if indstate==1
> , vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =      7,699
Group variable: cowcode                         Number of groups  =        148

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         17
                                                              avg =       52.0
                                                              max =         65

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(7)      =      15.57
Log pseudolikelihood  = -236.49565              Prob > chi2       =     0.0293

                                         (Std. Err. adjusted for 148 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
              leftistny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
          gedpartyautoc |   .5607227   .2773734    -1.17   0.242     .2126587    1.478472
            gedmilautoc |   1.244463   .6393433     0.43   0.670       .45465    3.406333
            gedmonautoc |   .9610481    .562433    -0.07   0.946     .3052091    3.026166
           gedpersautoc |   1.606716   .7111051     1.07   0.284     .6748569    3.825308
                  time1 |   .5108565   .5117237    -0.67   0.503      .071722    3.638693
                        |
        c.time1#c.time1 |   1.011495   .0140998     0.82   0.412      .984234    1.039511
                        |
c.time1#c.time1#c.time1 |    .999937   .0000634    -0.99   0.321     .9998127    1.000061
                        |
                  _cons |   2540.501   59795.73     0.33   0.739     2.35e-17    2.75e+23
------------------------+----------------------------------------------------------------
               /lnsig2u |  -.7267684     .83941                     -2.371982     .918445
------------------------+----------------------------------------------------------------
                sigma_u |   .6953192    .291829                      .3054434    1.582843
                    rho |    .227151   .1473613                      .0536728    .6036614
-----------------------------------------------------------------------------------------

. * Quadrature test
. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -236.49565     -236.49562     -236.49565
likelihood                   .00002598     -3.662e-07   Difference
                            -1.099e-07      1.548e-09   Relative difference
-----------------------------------------------------
leftistny:   -.57852871     -.57852871     -.57852904
gedpartyau~c                -2.998e-15     -3.382e-07   Difference
                             5.181e-15      5.847e-07   Relative difference
-----------------------------------------------------
leftistny:    .21870442      .21870442      .21870409
gedmilautoc                 -1.971e-15     -3.287e-07   Difference
                            -9.011e-15     -1.503e-06   Relative difference
-----------------------------------------------------
leftistny:    -.0397308      -.0397308     -.03973067
gedmonautoc                  2.630e-15      1.305e-07   Difference
                            -6.619e-14     -3.284e-06   Relative difference
-----------------------------------------------------
leftistny:    .47419219      .47419219      .47419166
gedpersautoc                -3.719e-15     -5.225e-07   Difference
                            -7.843e-15     -1.102e-06   Relative difference
-----------------------------------------------------
leftistny:   -.67166646     -.67166646     -.67166636
   time1                     8.882e-16      1.038e-07   Difference
                            -1.322e-15     -1.545e-07   Relative difference
-----------------------------------------------------
leftistny:    .01142936      .01142936      .01142936
c.time1#c.~1                -1.561e-17     -1.692e-09   Difference
                            -1.366e-15     -1.481e-07   Relative difference
-----------------------------------------------------
leftistny:   -.00006297     -.00006297     -.00006297
c.time1#c.~1                 8.132e-20      8.265e-12   Difference
                            -1.291e-15     -1.312e-07   Relative difference
-----------------------------------------------------
leftistny:    7.8401167      7.8401167      7.8401169
   _cons                     4.441e-15      2.545e-07   Difference
                             5.664e-16      3.247e-08   Relative difference
-----------------------------------------------------
lnsig2u:      -.7267684      -.7267684     -.72677692
   _cons                    -6.761e-14     -8.523e-06   Difference
                             9.303e-14      .00001173   Relative difference
-----------------------------------------------------

. * --Passed:  all coefficients change by less than .01 
. 
. * Figure 3.2. Years incumbent leader in power
. * Statistically significant relationship between yrsincleaderinpower and urban civic episodes
. * No statistically significant relationship between yrsincleaderinpower and social revolutionary episodes (vario
> us polynomial forms tested)
. * Urban civic
. xtcloglog urbancivicny yrsincleaderinpower c.time1##c.time1##c.time1 if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,661
Group variable: cowcode                         Number of groups  =        164

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
                                                              avg =       71.1
                                                              max =        115

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      32.16
Log pseudolikelihood  = -298.85435              Prob > chi2       =     0.0000

                                         (Std. Err. adjusted for 164 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
           urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
    yrsincleaderinpower |   1.059296     .01397     4.37   0.000     1.032266    1.087034
                  time1 |   .9782771   .0773353    -0.28   0.781     .8378615    1.142225
                        |
        c.time1#c.time1 |    1.00111   .0013404     0.83   0.407     .9984868    1.003741
                        |
c.time1#c.time1#c.time1 |   .9999944   6.73e-06    -0.83   0.408     .9999812    1.000008
                        |
                  _cons |   .0001371   .0002396    -5.09   0.000     4.46e-06    .0042146
------------------------+----------------------------------------------------------------
               /lnsig2u |  -.5252071   .6315522                     -1.763027    .7126124
------------------------+----------------------------------------------------------------
                sigma_u |   .7690467   .2428466                      .4141557    1.428045
                    rho |   .2644615   .1228505                      .0944282    .5535222
-----------------------------------------------------------------------------------------

. * Quadrature test
. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -298.85435     -298.85424     -298.85435
likelihood                   .00011281     -1.987e-06   Difference
                            -3.775e-07      6.648e-09   Relative difference
-----------------------------------------------------
urbancivicny: .05760461      .05760461      .05760447
yrsinclead~r                -4.440e-13     -1.330e-07   Difference
                            -7.708e-12     -2.309e-06   Relative difference
-----------------------------------------------------
urbancivicny: -.0219623      -.0219623      -.0219624
   time1                    -3.533e-13     -1.033e-07   Difference
                             1.609e-11      4.704e-06   Relative difference
-----------------------------------------------------
urbancivicny: .00110977      .00110977      .00110977
c.time1#c.~1                 3.960e-15      1.148e-09   Difference
                             3.568e-12      1.034e-06   Relative difference
-----------------------------------------------------
urbancivicny:-5.573e-06     -5.573e-06     -5.573e-06
c.time1#c.~1                -1.687e-17     -4.946e-12   Difference
                             3.028e-12      8.875e-07   Relative difference
-----------------------------------------------------
urbancivicny:-8.8951082     -8.8951082      -8.895098
   _cons                     3.252e-11      .00001018   Difference
                            -3.656e-12     -1.145e-06   Relative difference
-----------------------------------------------------
lnsig2u:     -.52520713     -.52520713     -.52522674
   _cons                    -5.790e-11     -.00001961   Difference
                             1.102e-10      .00003734   Relative difference
-----------------------------------------------------

. * --Passed:  all coefficients change by less than .01 
. * Social
. xtcloglog leftistny yrsincleaderinpower c.time1##c.time1##c.time1 if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,661
Group variable: cowcode                         Number of groups  =        164

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
                                                              avg =       71.1
                                                              max =        115

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      29.89
Log pseudolikelihood  = -379.30202              Prob > chi2       =     0.0000

                                         (Std. Err. adjusted for 164 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
              leftistny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
    yrsincleaderinpower |    1.00156   .0155939     0.10   0.920     .9714583    1.032595
                  time1 |      .9657    .030468    -1.11   0.269     .9077926    1.027301
                        |
        c.time1#c.time1 |   1.001435   .0007522     1.91   0.056      .999962    1.002911
                        |
c.time1#c.time1#c.time1 |   .9999868   4.89e-06    -2.71   0.007     .9999772    .9999964
                        |
                  _cons |   .0068245   .0028516   -11.94   0.000     .0030088    .0154789
------------------------+----------------------------------------------------------------
               /lnsig2u |  -.7040494   .6019727                     -1.883894    .4757955
------------------------+----------------------------------------------------------------
                sigma_u |   .7032627   .2116725                       .389868    1.268579
                    rho |   .2311641    .106987                       .084587     .494524
-----------------------------------------------------------------------------------------

. * Quadrature test
. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -379.30202     -379.30186     -379.30202
likelihood                   .00016133      2.277e-07   Difference
                            -4.253e-07     -6.003e-10   Relative difference
-----------------------------------------------------
leftistny:    .00155899      .00155899      .00155899
yrsinclead~r                -2.027e-14      2.425e-09   Difference
                            -1.300e-11      1.556e-06   Relative difference
-----------------------------------------------------
leftistny:   -.03490207     -.03490207     -.03490207
   time1                     1.934e-13      6.910e-09   Difference
                            -5.542e-12     -1.980e-07   Relative difference
-----------------------------------------------------
leftistny:    .00143424      .00143424      .00143424
c.time1#c.~1                -3.709e-15     -1.080e-10   Difference
                            -2.586e-12     -7.527e-08   Relative difference
-----------------------------------------------------
leftistny:   -.00001323     -.00001323     -.00001323
c.time1#c.~1                 1.383e-17      5.752e-14   Difference
                            -1.045e-12     -4.349e-09   Relative difference
-----------------------------------------------------
leftistny:   -4.9872362     -4.9872362     -4.9872335
   _cons                     2.996e-11      2.691e-06   Difference
                            -6.006e-12     -5.397e-07   Relative difference
-----------------------------------------------------
lnsig2u:     -.70404941     -.70404941     -.70405895
   _cons                    -8.871e-11     -9.542e-06   Difference
                             1.260e-10      .00001355   Relative difference
-----------------------------------------------------

. * --Passed:  all coefficients change by less than .01 
. * Testing polynomial forms
. xtcloglog leftistny c.yrsincleaderinpower##c.yrsincleaderinpower c.time1##c.time1##c.time1 if indstate==1, vce(r
> obust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,661
Group variable: cowcode                         Number of groups  =        164

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
                                                              avg =       71.1
                                                              max =        115

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(5)      =      34.59
Log pseudolikelihood  = -379.06874              Prob > chi2       =     0.0000

                                                             (Std. Err. adjusted for 164 clusters in cowcode)
-------------------------------------------------------------------------------------------------------------
                                            |               Robust
                                  leftistny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------------------------------+----------------------------------------------------------------
                        yrsincleaderinpower |   .9811257   .0284628    -0.66   0.511     .9268959    1.038528
                                            |
c.yrsincleaderinpower#c.yrsincleaderinpower |   1.000648   .0006204     1.04   0.296     .9994327    1.001865
                                            |
                                      time1 |   .9675876   .0311459    -1.02   0.306     .9084285    1.030599
                                            |
                            c.time1#c.time1 |   1.001421   .0007535     1.89   0.059      .999945    1.002899
                                            |
                    c.time1#c.time1#c.time1 |   .9999868   4.87e-06    -2.71   0.007     .9999772    .9999963
                                            |
                                      _cons |   .0067637   .0029014   -11.65   0.000     .0029178    .0156789
--------------------------------------------+----------------------------------------------------------------
                                   /lnsig2u |  -.6733001    .592526                      -1.83463    .4880294
--------------------------------------------+----------------------------------------------------------------
                                    sigma_u |   .7141587   .2115788                      .3995906    1.276363
                                        rho |   .2366742   .1070455                      .0884806    .4975823
-------------------------------------------------------------------------------------------------------------

. * Not statistically significant
. xtcloglog leftistny c.yrsincleaderinpower##c.yrsincleaderinpower##c.yrsincleaderinpower c.time1##c.time1##c.time
> 1 if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,661
Group variable: cowcode                         Number of groups  =        164

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
                                                              avg =       71.1
                                                              max =        115

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(6)      =      39.84
Log pseudolikelihood  = -377.90403              Prob > chi2       =     0.0000

                                                              (Std. Err. adjusted for 164 clusters in cowcode)
--------------------------------------------------------------------------------------------------------------
                                             |               Robust
                                   leftistny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
---------------------------------------------+----------------------------------------------------------------
                         yrsincleaderinpower |   .8900168   .0637121    -1.63   0.104     .7735078    1.024075
                                             |
 c.yrsincleaderinpower#c.yrsincleaderinpower |   1.007882   .0048397     1.64   0.102      .998441    1.017413
                                             |
 c.yrsincleaderinpower#c.yrsincleaderinpower#|
                       c.yrsincleaderinpower |   .9998825   .0000789    -1.49   0.136     .9997279    1.000037
                                             |
                                       time1 |      .9688   .0323907    -0.95   0.343     .9073507    1.034411
                                             |
                             c.time1#c.time1 |   1.001411   .0007719     1.83   0.067     .9998988    1.002924
                                             |
                     c.time1#c.time1#c.time1 |   .9999868   4.94e-06    -2.68   0.007     .9999771    .9999965
                                             |
                                       _cons |   .0076618   .0034678   -10.76   0.000     .0031555    .0186034
---------------------------------------------+----------------------------------------------------------------
                                    /lnsig2u |   -.674261   .6089088                       -1.8677    .5191783
---------------------------------------------+----------------------------------------------------------------
                                     sigma_u |   .7138157   .2173243                      .3930375    1.296397
                                         rho |   .2365006   .1099495                      .0858494    .5053693
--------------------------------------------------------------------------------------------------------------

. * Not statistically significant
. 
. * V-Dem executive corruption measure
. * Urban civic
. xtcloglog urbancivicny v2x_execorr c.time1##c.time1##c.time1 if indstate==1,  eform nolog vce(robust)

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,580
Group variable: cowcode                         Number of groups  =        161

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       71.9
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      29.38
Log pseudolikelihood  = -297.91275              Prob > chi2       =     0.0000

                                         (Std. Err. adjusted for 161 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
           urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
            v2x_execorr |   8.632152   5.003908     3.72   0.000     2.771356    26.88721
                  time1 |   .9615505   .0831322    -0.45   0.650     .8116714    1.139106
                        |
        c.time1#c.time1 |   1.001318   .0013954     0.95   0.344     .9985872    1.004057
                        |
c.time1#c.time1#c.time1 |   .9999936   6.88e-06    -0.93   0.351     .9999801    1.000007
                        |
                  _cons |   .0001329     .00027    -4.39   0.000     2.47e-06    .0071319
------------------------+----------------------------------------------------------------
               /lnsig2u |  -2.532605    2.80237                     -8.025149     2.95994
------------------------+----------------------------------------------------------------
                sigma_u |   .2818719   .3949547                      .0180868    4.392813
                    rho |   .0460754   .1231711                      .0001988    .9214519
-----------------------------------------------------------------------------------------

. * Quadrature test
. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -297.91275     -297.91275     -297.91275
likelihood                  -2.035e-10      4.820e-11   Difference
                             6.831e-13     -1.618e-13   Relative difference
-----------------------------------------------------
urbancivicny: 2.1554938      2.1554947      2.1554947
v2x_execorr                  8.719e-07      8.774e-07   Difference
                             4.045e-07      4.071e-07   Relative difference
-----------------------------------------------------
urbancivicny:-.03920817     -.03920816     -.03920816
   time1                     1.201e-08      1.209e-08   Difference
                            -3.062e-07     -3.083e-07   Relative difference
-----------------------------------------------------
urbancivicny: .00131743      .00131743      .00131743
c.time1#c.~1                -7.513e-11     -7.577e-11   Difference
                            -5.703e-08     -5.751e-08   Relative difference
-----------------------------------------------------
urbancivicny:-6.420e-06     -6.420e-06     -6.420e-06
c.time1#c.~1                 1.547e-13      1.566e-13   Difference
                            -2.410e-08     -2.439e-08   Relative difference
-----------------------------------------------------
urbancivicny:-8.9262364     -8.9262396     -8.9262396
   _cons                    -3.148e-06     -3.164e-06   Difference
                             3.527e-07      3.545e-07   Relative difference
-----------------------------------------------------
lnsig2u:     -2.5326048     -2.5325696     -2.5325693
   _cons                     .00003524      .00003549   Difference
                            -.00001391     -.00001401   Relative difference
-----------------------------------------------------

. * --Passed:  all coefficients change by less than .01 
. * Social
. xtcloglog leftistny v2x_execorr c.time1##c.time1##c.time1 if indstate==1,  eform nolog vce(robust)

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,580
Group variable: cowcode                         Number of groups  =        161

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       71.9
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      30.42
Log pseudolikelihood  = -378.06417              Prob > chi2       =     0.0000

                                         (Std. Err. adjusted for 161 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
              leftistny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
            v2x_execorr |   1.786318   .7119443     1.46   0.145     .8179168    3.901294
                  time1 |   .9557295   .0327221    -1.32   0.186     .8936998    1.022064
                        |
        c.time1#c.time1 |   1.001647   .0007747     2.13   0.033      1.00013    1.003166
                        |
c.time1#c.time1#c.time1 |   .9999855   4.95e-06    -2.93   0.003     .9999758    .9999952
                        |
                  _cons |   .0060524   .0035095    -8.81   0.000     .0019425     .018858
------------------------+----------------------------------------------------------------
               /lnsig2u |  -.7424384    .670923                     -2.057423    .5725464
------------------------+----------------------------------------------------------------
                sigma_u |   .6898927   .2314324                      .3574672    1.331456
                    rho |   .2244118   .1167749                       .072083    .5187028
-----------------------------------------------------------------------------------------

. * Quadrature test
. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -378.06417     -378.06404     -378.06417
likelihood                   .00013284      1.530e-07   Difference
                            -3.514e-07     -4.047e-10   Relative difference
-----------------------------------------------------
leftistny:    .58015678      .58015678        .580157
v2x_execorr                 -1.312e-11      2.179e-07   Difference
                            -2.262e-11      3.756e-07   Relative difference
-----------------------------------------------------
leftistny:   -.04528041     -.04528041      -.0452804
   time1                     8.364e-13      1.126e-08   Difference
                            -1.847e-11     -2.488e-07   Relative difference
-----------------------------------------------------
leftistny:    .00164549      .00164549      .00164549
c.time1#c.~1                -1.623e-14     -2.349e-10   Difference
                            -9.864e-12     -1.428e-07   Relative difference
-----------------------------------------------------
leftistny:   -.00001451     -.00001451     -.00001451
c.time1#c.~1                 6.741e-17      9.794e-13   Difference
                            -4.647e-12     -6.752e-08   Relative difference
-----------------------------------------------------
leftistny:   -5.1073016     -5.1073016     -5.1072998
   _cons                     1.234e-10      1.887e-06   Difference
                            -2.417e-11     -3.695e-07   Relative difference
-----------------------------------------------------
lnsig2u:     -.74243845     -.74243845     -.74244805
   _cons                    -3.944e-10     -9.602e-06   Difference
                             5.313e-10      .00001293   Relative difference
-----------------------------------------------------

. * --Passed:  all coefficients change by less than .01 
. 
. * Figure 3.3. GDP per capita
. * Urban civic 
. * Linear vs. Quadratic specification: quadratic is better (lower BIC and AIC)
. xtcloglog urbancivicny c.gdppcthl c.time1##c.time1##c.time1 if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,042
Group variable: cowcode                         Number of groups  =        164

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
                                                              avg =       67.3
                                                              max =        115

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      34.82
Log pseudolikelihood  = -304.05918              Prob > chi2       =     0.0000

                                         (Std. Err. adjusted for 164 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
           urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
               gdppcthl |    .937886   .0200535    -3.00   0.003     .8993941    .9780253
                  time1 |   .9387496   .0731789    -0.81   0.417     .8057412    1.093714
                        |
        c.time1#c.time1 |   1.001663   .0013059     1.27   0.202      .999107    1.004226
                        |
c.time1#c.time1#c.time1 |   .9999923   6.57e-06    -1.17   0.240     .9999794    1.000005
                        |
                  _cons |   .0008494   .0014368    -4.18   0.000     .0000308    .0233871
------------------------+----------------------------------------------------------------
               /lnsig2u |  -2.052453   1.975434                     -5.924232    1.819326
------------------------+----------------------------------------------------------------
                sigma_u |   .3583567   .3539549                      .0517094    2.483485
                    rho |   .0724162    .132694                      .0016229     .789452
-----------------------------------------------------------------------------------------

. estat ic

Akaike's information criterion and Bayesian information criterion

-----------------------------------------------------------------------------
       Model |        Obs  ll(null)  ll(model)      df         AIC        BIC
-------------+---------------------------------------------------------------
           . |     11,042         .  -304.0592       6    620.1184   663.9751
-----------------------------------------------------------------------------
               Note: N=Obs used in calculating BIC; see [R] BIC note.

. xtcloglog urbancivicny c.gdppcthl##c.gdppcthl c.time1##c.time1##c.time1 if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,042
Group variable: cowcode                         Number of groups  =        164

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
                                                              avg =       67.3
                                                              max =        115

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(5)      =      44.40
Log pseudolikelihood  = -298.49629              Prob > chi2       =     0.0000

                                         (Std. Err. adjusted for 164 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
           urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
               gdppcthl |   1.352199   .1456203     2.80   0.005     1.094898    1.669967
                        |
  c.gdppcthl#c.gdppcthl |   .9748752   .0073925    -3.36   0.001     .9604932    .9894725
                        |
                  time1 |   .9331617   .0731267    -0.88   0.377     .8003002     1.08808
                        |
        c.time1#c.time1 |   1.001773   .0013091     1.36   0.175     .9992101    1.004342
                        |
c.time1#c.time1#c.time1 |   .9999917   6.56e-06    -1.27   0.205     .9999788    1.000005
                        |
                  _cons |   .0004842   .0008109    -4.56   0.000     .0000182    .0128982
------------------------+----------------------------------------------------------------
               /lnsig2u |  -3.246525      6.543                     -16.07057    9.577519
------------------------+----------------------------------------------------------------
                sigma_u |   .1972541   .6453169                      .0003238    120.1522
                    rho |   .0231074   .1476979                      6.38e-08    .9998861
-----------------------------------------------------------------------------------------

. estat ic

Akaike's information criterion and Bayesian information criterion

-----------------------------------------------------------------------------
       Model |        Obs  ll(null)  ll(model)      df         AIC        BIC
-------------+---------------------------------------------------------------
           . |     11,042         .  -298.4963       7    610.9926   662.1588
-----------------------------------------------------------------------------
               Note: N=Obs used in calculating BIC; see [R] BIC note.

. * Quadrature test on quadratic
. quietly: xtcloglog urbancivicny c.gdppcthl##c.gdppcthl c.time1##c.time1##c.time1 if indstate==1, vce(robust) efo
> rm nolog

. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -298.49629     -298.49629     -298.49629
likelihood                   2.103e-11      2.188e-11   Difference
                            -7.046e-14     -7.332e-14   Relative difference
-----------------------------------------------------
urbancivicny: .30173245      .30173243      .30173243
gdppcthl                    -2.047e-08     -2.046e-08   Difference
                            -6.783e-08     -6.782e-08   Relative difference
-----------------------------------------------------
urbancivicny:-.02544586     -.02544586     -.02544586
c.gdppcthl~l                 1.848e-09      1.848e-09   Difference
                            -7.263e-08     -7.262e-08   Relative difference
-----------------------------------------------------
urbancivicny:-.06917681     -.06917682     -.06917682
   time1                    -3.088e-09     -3.088e-09   Difference
                             4.464e-08      4.464e-08   Relative difference
-----------------------------------------------------
urbancivicny: .00177103      .00177103      .00177103
c.time1#c.~1                 7.803e-12      7.802e-12   Difference
                             4.406e-09      4.405e-09   Relative difference
-----------------------------------------------------
urbancivicny:-8.325e-06     -8.325e-06     -8.325e-06
c.time1#c.~1                 5.977e-14      5.976e-14   Difference
                            -7.179e-09     -7.178e-09   Relative difference
-----------------------------------------------------
urbancivicny:-7.6330351      -7.633034      -7.633034
   _cons                     1.056e-06      1.056e-06   Difference
                            -1.384e-07     -1.384e-07   Relative difference
-----------------------------------------------------
lnsig2u:     -3.2465246     -3.2465819     -3.2465819
   _cons                    -.00005727     -.00005726   Difference
                             .00001764      .00001764   Relative difference
-----------------------------------------------------

. * --Passed:  all coefficients change by less than .01 
. * Social
. * Linear vs. quadratic specification: linear is better (lower BIC and AIC; quadratic not significant)
. xtcloglog leftistny gdppcthl c.time1##c.time1##c.time1 if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,042
Group variable: cowcode                         Number of groups  =        164

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
                                                              avg =       67.3
                                                              max =        115

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      41.83
Log pseudolikelihood  = -371.60028              Prob > chi2       =     0.0000

                                         (Std. Err. adjusted for 164 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
              leftistny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
               gdppcthl |   .8990926   .0422104    -2.27   0.023     .8200539    .9857493
                  time1 |   .9601783   .0310018    -1.26   0.208     .9012985    1.022905
                        |
        c.time1#c.time1 |   1.001404   .0007465     1.88   0.060     .9999422    1.002868
                        |
c.time1#c.time1#c.time1 |   .9999876   4.75e-06    -2.62   0.009     .9999783    .9999969
                        |
                  _cons |   .0140395   .0064059    -9.35   0.000     .0057407    .0343348
------------------------+----------------------------------------------------------------
               /lnsig2u |  -1.205998   .8547113                     -2.881202     .469205
------------------------+----------------------------------------------------------------
                sigma_u |   .5471681   .2338354                      .2367854    1.264406
                    rho |   .1539828    .111345                      .0329614    .4928767
-----------------------------------------------------------------------------------------

. estat ic

Akaike's information criterion and Bayesian information criterion

-----------------------------------------------------------------------------
       Model |        Obs  ll(null)  ll(model)      df         AIC        BIC
-------------+---------------------------------------------------------------
           . |     11,042         .  -371.6003       6    755.2006   799.0573
-----------------------------------------------------------------------------
               Note: N=Obs used in calculating BIC; see [R] BIC note.

. xtcloglog leftistny c.gdppcthl##c.gdppcthl c.time1##c.time1##c.time1 if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,042
Group variable: cowcode                         Number of groups  =        164

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
                                                              avg =       67.3
                                                              max =        115

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(5)      =      46.63
Log pseudolikelihood  = -371.22786              Prob > chi2       =     0.0000

                                         (Std. Err. adjusted for 164 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
              leftistny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
               gdppcthl |   1.010522    .127126     0.08   0.934     .7897004     1.29309
                        |
  c.gdppcthl#c.gdppcthl |   .9888158   .0098853    -1.13   0.261     .9696296    1.008382
                        |
                  time1 |   .9576359   .0311506    -1.33   0.183     .8984874    1.020678
                        |
        c.time1#c.time1 |   1.001465   .0007487     1.96   0.050     .9999987    1.002934
                        |
c.time1#c.time1#c.time1 |   .9999872   4.75e-06    -2.69   0.007     .9999779    .9999965
                        |
                  _cons |   .0120346   .0056612    -9.40   0.000     .0047865    .0302584
------------------------+----------------------------------------------------------------
               /lnsig2u |  -1.184168   .8603013                     -2.870328    .5019914
------------------------+----------------------------------------------------------------
                sigma_u |   .5531732   .2379478                      .2380764    1.285305
                    rho |   .1568482   .1137721                      .0333098    .5010728
-----------------------------------------------------------------------------------------

. estat ic

Akaike's information criterion and Bayesian information criterion

-----------------------------------------------------------------------------
       Model |        Obs  ll(null)  ll(model)      df         AIC        BIC
-------------+---------------------------------------------------------------
           . |     11,042         .  -371.2279       7    756.4557   807.6219
-----------------------------------------------------------------------------
               Note: N=Obs used in calculating BIC; see [R] BIC note.

. * Quadrature test on linear
. quietly: xtcloglog leftistny gdppcthl c.time1##c.time1##c.time1 if indstate==1, vce(robust) eform nolog

. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -371.60028     -371.60027     -371.60028
likelihood                   7.670e-06     -4.093e-08   Difference
                            -2.064e-08      1.102e-10   Relative difference
-----------------------------------------------------
leftistny:   -.10636923     -.10636923     -.10636926
gdppcthl                    -2.744e-13     -2.974e-08   Difference
                             2.580e-12      2.796e-07   Relative difference
-----------------------------------------------------
leftistny:   -.04063633     -.04063633     -.04063633
   time1                     5.266e-14      1.707e-09   Difference
                            -1.296e-12     -4.200e-08   Relative difference
-----------------------------------------------------
leftistny:    .00140325      .00140325      .00140325
c.time1#c.~1                -1.463e-15     -4.984e-11   Difference
                            -1.042e-12     -3.552e-08   Relative difference
-----------------------------------------------------
leftistny:   -.00001244     -.00001244     -.00001244
c.time1#c.~1                 8.226e-18      3.149e-13   Difference
                            -6.614e-13     -2.531e-08   Relative difference
-----------------------------------------------------
leftistny:   -4.2658807     -4.2658807     -4.2658803
   _cons                     8.264e-12      3.952e-07   Difference
                            -1.937e-12     -9.263e-08   Relative difference
-----------------------------------------------------
lnsig2u:     -1.2059984     -1.2059984     -1.2060038
   _cons                    -4.271e-11     -5.462e-06   Difference
                             3.542e-11      4.529e-06   Relative difference
-----------------------------------------------------

. * --Passed:  all coefficients change by less than .01 
. 
. * Figure 3.4. Economic growth
. * Urban civic
. xtcloglog urbancivicny gdppcgrow1yrl c.time1##c.time1##c.time1 if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,980
Group variable: cowcode                         Number of groups  =        164

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          9
                                                              avg =       67.0
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      33.01
Log pseudolikelihood  = -306.04863              Prob > chi2       =     0.0000

                                         (Std. Err. adjusted for 164 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
           urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
          gdppcgrow1yrl |   1.023162   .0099809     2.35   0.019     1.003786    1.042912
                  time1 |   .9220375   .0741837    -1.01   0.313     .7875245    1.079526
                        |
        c.time1#c.time1 |   1.001959   .0013547     1.45   0.148      .999307    1.004617
                        |
c.time1#c.time1#c.time1 |   .9999907   6.81e-06    -1.37   0.171     .9999773    1.000004
                        |
                  _cons |   .0009143   .0015393    -4.16   0.000     .0000337    .0247856
------------------------+----------------------------------------------------------------
               /lnsig2u |  -1.796819   1.469043                      -4.67609    1.082451
------------------------+----------------------------------------------------------------
                sigma_u |   .4072168   .2991094                      .0965162    1.718111
                    rho |   .0915779   .1222117                      .0056312    .6421599
-----------------------------------------------------------------------------------------

. * Quadrature test
. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -306.04863     -306.04863     -306.04863
likelihood                  -3.231e-08      1.091e-11   Difference
                             1.056e-10     -3.566e-14   Relative difference
-----------------------------------------------------
urbancivicny: .02289775      .02289775      .02289775
gdppcgrow1~l                -4.921e-11     -2.824e-11   Difference
                            -2.149e-09     -1.234e-09   Relative difference
-----------------------------------------------------
urbancivicny:-.08116934     -.08116934     -.08116934
   time1                    -7.987e-10     -4.001e-10   Difference
                             9.840e-09      4.930e-09   Relative difference
-----------------------------------------------------
urbancivicny: .00195679      .00195679      .00195679
c.time1#c.~1                -1.181e-11     -4.907e-12   Difference
                            -6.035e-09     -2.508e-09   Relative difference
-----------------------------------------------------
urbancivicny:-9.333e-06     -9.333e-06     -9.333e-06
c.time1#c.~1                 1.030e-13      4.494e-14   Difference
                            -1.103e-08     -4.815e-09   Relative difference
-----------------------------------------------------
urbancivicny:-6.9974042     -6.9974035     -6.9974039
   _cons                     7.329e-07      3.401e-07   Difference
                            -1.047e-07     -4.861e-08   Relative difference
-----------------------------------------------------
lnsig2u:     -1.7968192     -1.7968282     -1.7968239
   _cons                    -8.985e-06     -4.705e-06   Difference
                             5.001e-06      2.619e-06   Relative difference
-----------------------------------------------------

. * --Passed:  all coefficients change by less than .01 
. * Social
. xtcloglog leftistny gdppcgrow1yrl c.time1##c.time1##c.time1 if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,980
Group variable: cowcode                         Number of groups  =        164

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          9
                                                              avg =       67.0
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      49.72
Log pseudolikelihood  = -368.21201              Prob > chi2       =     0.0000

                                         (Std. Err. adjusted for 164 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
              leftistny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
          gdppcgrow1yrl |   .9457661   .0112912    -4.67   0.000     .9238926    .9681574
                  time1 |   .9412079   .0323814    -1.76   0.078      .879834    1.006863
                        |
        c.time1#c.time1 |    1.00184   .0007972     2.31   0.021     1.000279    1.003404
                        |
c.time1#c.time1#c.time1 |   .9999845   5.18e-06    -2.99   0.003     .9999743    .9999947
                        |
                  _cons |   .0140165   .0067618    -8.85   0.000     .0054451    .0360805
------------------------+----------------------------------------------------------------
               /lnsig2u |  -.8113529   .6521318                     -2.089508     .466802
------------------------+----------------------------------------------------------------
                sigma_u |   .6665258   .2173313                      .3517784    1.262888
                    rho |   .2126453   .1091847                      .0699662     .492276
-----------------------------------------------------------------------------------------

. * Quadrature test
. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -368.21201     -368.21192     -368.21201
likelihood                   .00009483     -1.722e-08   Difference
                            -2.575e-07      4.677e-11   Relative difference
-----------------------------------------------------
leftistny:   -.05576004     -.05576004     -.05576003
gdppcgrow1~l                 7.999e-14      1.092e-08   Difference
                            -1.435e-12     -1.959e-07   Relative difference
-----------------------------------------------------
leftistny:   -.06059118     -.06059118     -.06059117
   time1                     2.612e-13      9.489e-09   Difference
                            -4.311e-12     -1.566e-07   Relative difference
-----------------------------------------------------
leftistny:     .0018384       .0018384       .0018384
c.time1#c.~1                -5.243e-15     -2.088e-10   Difference
                            -2.852e-12     -1.136e-07   Relative difference
-----------------------------------------------------
leftistny:   -.00001549     -.00001549     -.00001549
c.time1#c.~1                 2.408e-17      8.035e-13   Difference
                            -1.554e-12     -5.186e-08   Relative difference
-----------------------------------------------------
leftistny:   -4.2675215     -4.2675215     -4.2675192
   _cons                     2.660e-11      2.241e-06   Difference
                            -6.234e-12     -5.252e-07   Relative difference
-----------------------------------------------------
lnsig2u:     -.81135288     -.81135288     -.81136457
   _cons                    -1.027e-10     -.00001169   Difference
                             1.265e-10      .00001441   Relative difference
-----------------------------------------------------

. * --Passed:  all coefficients change by less than .01 
. 
. * Absence of relationship of economic growth to urban civic revolution in upper-income countries
. *       More consistent relationship among lower middle-income countries
. xtcloglog urbancivicny i.gdppcquartersl##c.gdppcgrow1yrl c.time1##c.time1##c.time1 if indstate==1, vce(robust) e
> form nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,980
Group variable: cowcode                         Number of groups  =        164

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          9
                                                              avg =       67.0
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      72.92
Log pseudolikelihood  = -298.23054              Prob > chi2       =     0.0000

                                                (Std. Err. adjusted for 164 clusters in cowcode)
------------------------------------------------------------------------------------------------
                               |               Robust
                  urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------------------------+----------------------------------------------------------------
                gdppcquartersl |
                            2  |   1.135674    .319499     0.45   0.651     .6543124    1.971161
                            3  |   1.120237   .5574428     0.23   0.820     .4224181    2.970828
                            4  |   .1173629   .1237091    -2.03   0.042     .0148697    .9263167
                               |
                 gdppcgrow1yrl |   .9775867   .0154451    -1.43   0.151     .9477787    1.008332
                               |
gdppcquartersl#c.gdppcgrow1yrl |
                            2  |   1.051118    .020754     2.52   0.012     1.011218    1.092592
                            3  |   1.089363   .0498515     1.87   0.061     .9959091    1.191586
                            4  |   .9735261   .0193323    -1.35   0.177     .9363634    1.012164
                               |
                         time1 |   .9216375   .0751456    -1.00   0.317     .7855206    1.081341
                               |
               c.time1#c.time1 |   1.001971   .0013681     1.44   0.149     .9992927    1.004656
                               |
       c.time1#c.time1#c.time1 |   .9999907   6.87e-06    -1.35   0.177     .9999773    1.000004
                               |
                         _cons |   .0009348   .0015908    -4.10   0.000     .0000333    .0262592
-------------------------------+----------------------------------------------------------------
                      /lnsig2u |  -3.572014   8.848644                     -20.91504    13.77101
-------------------------------+----------------------------------------------------------------
                       sigma_u |   .1676282   .7416411                      .0000287    977.9951
                           rho |   .0167954   .1461202                      5.02e-10    .9999983
------------------------------------------------------------------------------------------------

. * Quadrature test
. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -298.23054     -298.23054     -298.23054
likelihood                   1.023e-12      6.253e-13   Difference
                            -3.431e-15     -2.097e-15   Relative difference
-----------------------------------------------------
urbancivicny: .12722613       .1272261       .1272261
2.gdppcqua~l                -2.561e-08     -2.277e-08   Difference
                            -2.013e-07     -1.790e-07   Relative difference
-----------------------------------------------------
urbancivicny: .11354054      .11354049       .1135405
3.gdppcqua~l                -4.746e-08     -4.219e-08   Difference
                            -4.180e-07     -3.716e-07   Relative difference
-----------------------------------------------------
urbancivicny:-2.1424843     -2.1424843     -2.1424843
4.gdppcqua~l                 9.850e-09      8.756e-09   Difference
                            -4.598e-09     -4.087e-09   Relative difference
-----------------------------------------------------
urbancivicny:-.02266826     -.02266826     -.02266826
gdppcgrow1~l                -3.433e-09     -3.051e-09   Difference
                             1.514e-07      1.346e-07   Relative difference
-----------------------------------------------------
urbancivicny: .04985406      .04985407      .04985407
2.gdppcqua~l                 5.302e-09      4.713e-09   Difference
                             1.063e-07      9.453e-08   Relative difference
-----------------------------------------------------
urbancivicny: .08559276      .08559276      .08559276
3.gdppcqua~l                 5.534e-09      4.919e-09   Difference
                             6.465e-08      5.747e-08   Relative difference
-----------------------------------------------------
urbancivicny:-.02683063     -.02683063     -.02683063
4.gdppcqua~l                 4.657e-10      4.139e-10   Difference
                            -1.736e-08     -1.543e-08   Relative difference
-----------------------------------------------------
urbancivicny:-.08160327     -.08160326     -.08160326
   time1                     1.991e-09      1.770e-09   Difference
                            -2.440e-08     -2.169e-08   Relative difference
-----------------------------------------------------
urbancivicny: .00196857      .00196857      .00196857
c.time1#c.~1                -1.197e-11     -1.064e-11   Difference
                            -6.081e-09     -5.406e-09   Relative difference
-----------------------------------------------------
urbancivicny:-9.275e-06     -9.275e-06     -9.275e-06
c.time1#c.~1                 1.714e-14      1.524e-14   Difference
                            -1.848e-09     -1.643e-09   Relative difference
-----------------------------------------------------
urbancivicny:-6.9752198     -6.9752202     -6.9752201
   _cons                    -3.880e-07     -3.449e-07   Difference
                             5.563e-08      4.945e-08   Relative difference
-----------------------------------------------------
lnsig2u:     -3.5720137     -3.5719957     -3.5719977
   _cons                     .00001791      .00001592   Difference
                            -5.015e-06     -4.458e-06   Relative difference
-----------------------------------------------------

. * --Passed:  all coefficients change by less than .01 
. 
. * Lack of bivariate relationship between oil production and revolution (either social or urban civic)
. * Urban civic
. xtcloglog urbancivicny lnoill c.time1##c.time1##c.time1 if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,560
Group variable: cowcode                         Number of groups  =        162

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         22
                                                              avg =       71.4
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      24.91
Log pseudolikelihood  = -306.89631              Prob > chi2       =     0.0001

                                         (Std. Err. adjusted for 162 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
           urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
                 lnoill |   1.010538    .030809     0.34   0.731     .9519217    1.072763
                  time1 |   .9444761   .0792696    -0.68   0.496     .8012164    1.113351
                        |
        c.time1#c.time1 |   1.001622   .0013847     1.17   0.241     .9989115     1.00434
                        |
c.time1#c.time1#c.time1 |   .9999922   6.89e-06    -1.14   0.256     .9999787    1.000006
                        |
                  _cons |   .0005282   .0009415    -4.23   0.000     .0000161    .0173779
------------------------+----------------------------------------------------------------
               /lnsig2u |  -1.737055   1.410395                     -4.501378    1.027269
------------------------+----------------------------------------------------------------
                sigma_u |    .419569    .295879                      .1053266    1.671354
                    rho |   .0966726   .1231656                       .006699    .6293824
-----------------------------------------------------------------------------------------

. * Quadrature test
. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -306.89631     -306.89631     -306.89631
likelihood                  -4.949e-08      2.268e-11   Difference
                             1.612e-10     -7.390e-14   Relative difference
-----------------------------------------------------
urbancivicny: .01048243      .01048244      .01048244
  lnoill                     9.593e-09      4.582e-09   Difference
                             9.152e-07      4.372e-07   Relative difference
-----------------------------------------------------
urbancivicny:-.05712494     -.05712493     -.05712493
   time1                     7.770e-09      3.636e-09   Difference
                            -1.360e-07     -6.365e-08   Relative difference
-----------------------------------------------------
urbancivicny: .00162057      .00162057      .00162057
c.time1#c.~1                -1.433e-10     -6.738e-11   Difference
                            -8.842e-08     -4.158e-08   Relative difference
-----------------------------------------------------
urbancivicny:-7.834e-06     -7.834e-06     -7.834e-06
c.time1#c.~1                 6.975e-13      3.286e-13   Difference
                            -8.904e-08     -4.195e-08   Relative difference
-----------------------------------------------------
urbancivicny:-7.5459413     -7.5459404     -7.5459408
   _cons                     8.902e-07      4.279e-07   Difference
                            -1.180e-07     -5.671e-08   Relative difference
-----------------------------------------------------
lnsig2u:     -1.7370548     -1.7370679     -1.7370618
   _cons                    -.00001312     -7.079e-06   Difference
                             7.551e-06      4.075e-06   Relative difference
-----------------------------------------------------

. * --Passed:  all coefficients change by less than .01 
. * Social
. xtcloglog leftistny lnoill c.time1##c.time1##c.time1 if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,560
Group variable: cowcode                         Number of groups  =        162

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         22
                                                              avg =       71.4
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      30.60
Log pseudolikelihood  =  -371.9194              Prob > chi2       =     0.0000

                                         (Std. Err. adjusted for 162 clusters in cowcode)
-----------------------------------------------------------------------------------------
                        |               Robust
              leftistny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
------------------------+----------------------------------------------------------------
                 lnoill |   1.057081   .0398787     1.47   0.141     .9817396    1.138204
                  time1 |   .9528851   .0334684    -1.37   0.169     .8894951    1.020793
                        |
        c.time1#c.time1 |   1.001618   .0007878     2.06   0.040     1.000075    1.003163
                        |
c.time1#c.time1#c.time1 |   .9999858   5.01e-06    -2.83   0.005      .999976    .9999956
                        |
                  _cons |   .0080904   .0039277    -9.92   0.000     .0031241    .0209512
------------------------+----------------------------------------------------------------
               /lnsig2u |  -.7691891   .6475971                     -2.038456    .5000778
------------------------+----------------------------------------------------------------
                sigma_u |   .6807266   .2204183                      .3608734    1.284075
                    rho |   .2197902   .1110515                       .073362    .5005944
-----------------------------------------------------------------------------------------

. * Quadrature test
. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log           -371.9194     -371.91928      -371.9194
likelihood                   .00011788      1.690e-07   Difference
                            -3.169e-07     -4.543e-10   Relative difference
-----------------------------------------------------
leftistny:    .05551105      .05551105       .0555111
  lnoill                    -4.145e-14      5.106e-08   Difference
                            -7.466e-13      9.198e-07   Relative difference
-----------------------------------------------------
leftistny:    -.0482609      -.0482609     -.04826089
   time1                     7.131e-14      7.883e-09   Difference
                            -1.478e-12     -1.634e-07   Relative difference
-----------------------------------------------------
leftistny:    .00161672      .00161672      .00161672
c.time1#c.~1                -9.886e-16     -1.409e-10   Difference
                            -6.115e-13     -8.714e-08   Relative difference
-----------------------------------------------------
leftistny:    -.0000142      -.0000142      -.0000142
c.time1#c.~1                 2.428e-18      4.144e-13   Difference
                            -1.709e-13     -2.918e-08   Relative difference
-----------------------------------------------------
leftistny:   -4.8170772     -4.8170772     -4.8170756
   _cons                     1.186e-11      1.581e-06   Difference
                            -2.463e-12     -3.281e-07   Relative difference
-----------------------------------------------------
lnsig2u:     -.76918914     -.76918914     -.76920122
   _cons                    -4.463e-11     -.00001208   Difference
                             5.802e-11       .0000157   Relative difference
-----------------------------------------------------

. * --Passed:  all coefficients change by less than .01 
. 
. 
. * =============================================================================
. * ROBUSTNESS TESTS FOR URBAN CIVIC REVOLUTIONARY EPISODES MODEL  (TABLE 3.1)
. * =============================================================================
. * ++++++++++++++++++++++++++++++++++++++++++++++++++++++++
. * Robustness test for possible issues of multicollinearity
. * ++++++++++++++++++++++++++++++++++++++++++++++++++++++++
. * Checking for possible multicollinearity using variance inflation factors, with variables in Model 4 in Table 3
> .1
. quietly: reg urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill post
> coldwar if indstate==1

. estat vif

    Variable |       VIF       1/VIF  
-------------+----------------------
    gdppcthl |     13.23    0.075609
   gdppcthl2 |      9.61    0.104008
     polityl |      1.98    0.505850
 v2x_execorr |      1.94    0.516175
      lnoill |      1.60    0.623168
    polityl2 |      1.49    0.671228
 postcoldwar |      1.28    0.779371
      lnpopl |      1.26    0.796809
yrsinclead~r |      1.24    0.804525
-------------+----------------------
    Mean VIF |      3.74

. * RESULT:  GDP per capita and its quadratic term are potentially problematic,
. *       as VIF approaches or is greater than 10
. *  To test this whetheer multicollinearity affected the results, as recommended
. *       in some sources, I centered the variable and re-ran the regression to see 
. *       if centering the variable changed any parameters of the model
. sum gdppcthl if indstate==1

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
    gdppcthl |     11,042    4.658525    5.382156     .20654   42.91624

. local gdppcmean = r(mean)

. generate centeredgdppcthl= gdppcthl - `gdppcmean' if indstate==1
(7,347 missing values generated)

. generate centeredgdppcthl2 = centeredgdppcthl * centeredgdppcthl if indstate==1
(7,347 missing values generated)

. quietly: reg urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill post
> coldwar if indstate==1

. estat vif

    Variable |       VIF       1/VIF  
-------------+----------------------
    gdppcthl |     13.23    0.075609
   gdppcthl2 |      9.61    0.104008
     polityl |      1.98    0.505850
 v2x_execorr |      1.94    0.516175
      lnoill |      1.60    0.623168
    polityl2 |      1.49    0.671228
 postcoldwar |      1.28    0.779371
      lnpopl |      1.26    0.796809
yrsinclead~r |      1.24    0.804525
-------------+----------------------
    Mean VIF |      3.74

. * New gdppcthl variables now have variance inflation factor within reasonable limits
. * Then checked to see if it made a difference in the regression outcomes
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 gdppcgrow1yrl polityl polityl2 yrsincleaderinpower v2x_execorr 
> lnoill postcoldwar if indstate==1, eform nolog

Random-effects complementary log-log model      Number of obs     =     10,494
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       66.8
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      83.36
Log likelihood  = -263.48094                    Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.508058   .1639838     3.78   0.000     1.218596    1.866278
           gdppcthl |   1.805861   .2720535     3.92   0.000     1.344158    2.426152
          gdppcthl2 |   .9678286    .011487    -2.76   0.006     .9455743    .9906067
      gdppcgrow1yrl |   1.011886   .0227608     0.53   0.599     .9682448    1.057494
            polityl |   .9271115    .027432    -2.56   0.011     .8748751    .9824669
           polityl2 |   .9842485   .0055676    -2.81   0.005     .9733964    .9952216
yrsincleaderinpower |    1.03299   .0142368     2.36   0.019      1.00546    1.061274
        v2x_execorr |   5.073153   3.401343     2.42   0.015      1.36327    18.87878
             lnoill |   .8794047   .0319671    -3.54   0.000     .8189302    .9443451
        postcoldwar |   6.911917   2.977811     4.49   0.000     2.970855    16.08109
              _cons |   6.13e-06   7.82e-06    -9.42   0.000     5.04e-07    .0000746
--------------------+----------------------------------------------------------------
           /lnsig2u |  -10.56401   25.71614                     -60.96672    39.83869
--------------------+----------------------------------------------------------------
            sigma_u |   .0050822   .0653475                      5.77e-14    4.48e+08
                rho |   .0000157   .0004038                      2.02e-27           1
-------------------------------------------------------------------------------------
LR test of rho=0: chibar2(01) = 5.8e-05                Prob >= chibar2 = 0.497

. estimates store mod1

. xtcloglog urbancivicny lnpopl centeredgdppcthl centeredgdppcthl2 gdppcgrow1yrl polityl polityl2 yrsincleaderinpo
> wer v2x_execorr lnoill postcoldwar if indstate==1, eform nolog

Random-effects complementary log-log model      Number of obs     =     10,494
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       66.8
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      83.36
Log likelihood  = -263.48094                    Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.508058   .1639838     3.78   0.000     1.218596    1.866278
   centeredgdppcthl |   1.331582   .0830366     4.59   0.000     1.178386    1.504694
  centeredgdppcthl2 |   .9678287    .011487    -2.76   0.006     .9455743    .9906067
      gdppcgrow1yrl |   1.011886   .0227608     0.53   0.599     .9682448    1.057494
            polityl |   .9271115    .027432    -2.56   0.011     .8748751    .9824669
           polityl2 |   .9842485   .0055676    -2.81   0.005     .9733964    .9952216
yrsincleaderinpower |    1.03299   .0142368     2.36   0.019      1.00546    1.061274
        v2x_execorr |   5.073152   3.401343     2.42   0.015      1.36327    18.87878
             lnoill |   .8794047   .0319671    -3.54   0.000     .8189302    .9443451
        postcoldwar |   6.911917   2.977811     4.49   0.000     2.970855    16.08109
              _cons |   .0000473   .0000576    -8.18   0.000     4.36e-06    .0005139
--------------------+----------------------------------------------------------------
           /lnsig2u |  -10.56401   25.71614                     -60.96672    39.83869
--------------------+----------------------------------------------------------------
            sigma_u |   .0050822   .0653475                      5.77e-14    4.48e+08
                rho |   .0000157   .0004038                      2.02e-27           1
-------------------------------------------------------------------------------------
LR test of rho=0: chibar2(01) = 5.8e-05                Prob >= chibar2 = 0.497

. estimates store mod2

. * RESULT: No difference in patterns of statistical significance on visual inspection
. * Hausman test of the two models
. hausman mod1 mod2

                 ---- Coefficients ----
             |      (b)          (B)            (b-B)     sqrt(diag(V_b-V_B))
             |      mod1         mod2        Difference          S.E.
-------------+----------------------------------------------------------------
      lnpopl |    .4108228     .4108228        1.15e-09               .
gdppcgrow1~l |     .011816      .011816       -1.66e-10               .
     polityl |   -.0756814    -.0756814       -6.29e-11        9.34e-07
    polityl2 |   -.0158769    -.0158769       -6.97e-11        1.63e-07
yrsinclead~r |    .0324576     .0324576       -1.06e-10        3.26e-07
 v2x_execorr |    1.623962     1.623962        1.42e-08        .0000328
      lnoill |     -.12851      -.12851       -1.05e-09        8.29e-07
 postcoldwar |    1.933247     1.933247        1.90e-09               .
------------------------------------------------------------------------------
                       b = consistent under Ho and Ha; obtained from xtcloglog
        B = inconsistent under Ha, efficient under Ho; obtained from xtcloglog

    Test:  Ho:  difference in coefficients not systematic

                  chi2(8) = (b-B)'[(V_b-V_B)^(-1)](b-B)
                          =        0.00
                Prob>chi2 =      1.0000
                (V_b-V_B is not positive definite)

. * RESULT:  No difference in the regression coefficients of the two models: can 
. *       safely keep gdppcthl and gdppcthl2 in the specification
. drop _est_mod1 _est_mod2

. macro drop _all

. drop centeredgdppcthl centeredgdppcthl2

. 
. * +++++++++++++++++++++++++++
. * Bootstrap country-clusters 
. * +++++++++++++++++++++++++++
. * WARNING:  CAN TAKE A WHILE TO COMPUTE
. * Must use pooled model, with clustered standard errors
. * Warning:  will be reloading dataset after this command
. clear

. use revspredictbycntryyr, clear

. xtset, clear

. bootstrap , reps(1000) cluster(cowcode) idcluster(newid) : cloglog urbancivicny lnpopl gdppcthl gdppcthl2 polity
> l polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldwar if indstate==1, vce(cluster cowcode) eform
(running cloglog on estimation sample)

Bootstrap replications (1000)
----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5 
..................................................    50
..................................................   100
..................................................   150
..................................................   200
..................................................   250
..................................................   300
..................................................   350
..................................................   400
.................x................................   450
..................................................   500
..................................................   550
..................................................   600
..................................................   650
..................................................   700
..................................................   750
..................................................   800
..................................................   850
..................................................   900
..................................................   950
..................................................  1000

Complementary log-log regression                Number of obs     =     10,516
                                                Zero outcomes     =     10,466
                                                Nonzero outcomes  =         50

                                                Wald chi2(9)      =      75.44
Log pseudolikelihood = -263.63106               Prob > chi2       =     0.0000

                                      (Replications based on 157 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |   Observed   Bootstrap                         Normal-based
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.518912   .1692926     3.75   0.000     1.220845    1.889751
           gdppcthl |   1.824335   .2693241     4.07   0.000     1.365976      2.4365
          gdppcthl2 |   .9673957   .0107368    -2.99   0.003     .9465792    .9886699
            polityl |   .9269707    .035077    -2.00   0.045     .8607086    .9983341
           polityl2 |   .9841294   .0063199    -2.49   0.013     .9718203    .9965944
yrsincleaderinpower |    1.03303   .0127873     2.63   0.009     1.008269    1.058399
        v2x_execorr |   5.163875   4.268555     1.99   0.047     1.021778    26.09725
             lnoill |   .8778001   .0285184    -4.01   0.000     .8236474    .9355132
        postcoldwar |   6.900119   3.170393     4.20   0.000     2.803855    16.98078
              _cons |   5.73e-06   8.36e-06    -8.27   0.000     3.28e-07    .0001001
-------------------------------------------------------------------------------------
Note: One or more parameters could not be estimated in 1 bootstrap replicate;
      standard-error estimates include only complete replications.

. estat bootstrap, eform all

Complementary log-log regression                Number of obs     =     10,516
                                                Replications      =        999

                               (Replications based on 157 clusters in cowcode)
------------------------------------------------------------------------------
             |    Observed               Bootstrap
urbancivicny |      exp(b)       Bias    Std. Err.  [95% Conf. Interval]
-------------+----------------------------------------------------------------
      lnpopl |   1.5189117   .0062957   .16929261    1.220845   1.889751   (N)
             |                                       1.234494   1.905959   (P)
             |                                       1.222955   1.897449  (BC)
    gdppcthl |   1.8243353   .0515036   .26932415    1.365976     2.4365   (N)
             |                                       1.443315   2.587939   (P)
             |                                       1.381965   2.468799  (BC)
   gdppcthl2 |   .96739568  -.0027637   .01073679    .9465792   .9886699   (N)
             |                                       .9403872   .9815044   (P)
             |                                       .9434382   .9824724  (BC)
     polityl |   .92697075  -.0052524   .03507704    .8607086   .9983341   (N)
             |                                       .8520005    .987633   (P)
             |                                       .8589333   .9949931  (BC)
    polityl2 |   .98412937  -.0005179   .00631986    .9718203   .9965944   (N)
             |                                       .9701071   .9967226   (P)
             |                                       .9714537   .9972732  (BC)
yrsinclead~r |   1.0330301   .0014379    .0127873    1.008269   1.058399   (N)
             |                                       1.011081   1.060986   (P)
             |                                       1.009744   1.058194  (BC)
 v2x_execorr |   5.1638749   .0841462   4.2685546    1.021778   26.09725   (N)
             |                                       1.237899   29.03565   (P)
             |                                       1.242753   29.03565  (BC)
      lnoill |    .8778001  -.0031151   .02851844    .8236474   .9355132   (N)
             |                                       .8192704   .9350024   (P)
             |                                       .8247999    .942234  (BC)
 postcoldwar |   6.9001189   .4122237   3.1703928    2.803855   16.98078   (N)
             |                                       3.137168    19.3746   (P)
             |                                       3.041644   17.49635  (BC)
------------------------------------------------------------------------------
(N)    normal confidence interval
(P)    percentile confidence interval
(BC)   bias-corrected confidence interval
Note: One or more parameters could not be estimated in 1 bootstrap replicate;
      standard-error estimates include only complete replications.

. *  Result:  no changes in signs or patterns of statistical significance
. clear

. use revspredictbycntryyr.dta

. 
. * ++++++++++++++++++++++++++++
. * Other estimation techniques
. * ++++++++++++++++++++++++++++
. * Rare event framework for Model 4
. * probability of an urban civic revolt across all cases in the sample (.005) taken as the pc parameter
. relogit urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldw
> ar if indstate==1, cluster(cowcode) pc (.0048)
(7,786 missing values generated)


Corrected logit estimates                             Number of obs =    10516

-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   .4209999   .0980301     4.29   0.000     .2288644    .6131354
           gdppcthl |   .5640568   .1312309     4.30   0.000      .306849    .8212647
          gdppcthl2 |  -.0291225   .0092573    -3.15   0.002    -.0472664   -.0109785
            polityl |  -.0740166   .0322446    -2.30   0.022    -.1372148   -.0108184
           polityl2 |  -.0154494   .0056158    -2.75   0.006     -.026456   -.0044427
yrsincleaderinpower |    .034092   .0114406     2.98   0.003     .0116687    .0565152
        v2x_execorr |   1.577716   .8003907     1.97   0.049     .0089788    3.146453
             lnoill |  -.1296778   .0304235    -4.26   0.000    -.1893067   -.0700489
        postcoldwar |   1.874305   .4183023     4.48   0.000     1.054448    2.694163
              _cons |  -11.88128   1.370856    -8.67   0.000    -14.56811    -9.19445
-------------------------------------------------------------------------------------

. *       RESULT: all findings remain unchanged
. 
. * Population-averaged complementary log-log panel framework with different correlation structures for Model 4
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar time1 timesq timecub if  indstate==1, pa corr(exc) eform nolog vce(robust)

GEE population-averaged model                   Number of obs     =     10,516
Group variable:                    cowcode      Number of groups  =        157
Link:                              cloglog      Obs per group:
Family:                           binomial                    min =         21
Correlation:                  exchangeable                    avg =       67.0
                                                              max =        114
                                                Wald chi2(12)     =     105.68
Scale parameter:                         1      Prob > chi2       =     0.0000

                                       (Std. Err. adjusted for clustering on cowcode)
-------------------------------------------------------------------------------------
                    |             Semirobust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.539009   .1586731     4.18   0.000     1.257424    1.883653
           gdppcthl |   1.802543    .229785     4.62   0.000     1.404028     2.31417
          gdppcthl2 |   .9683197   .0089517    -3.48   0.000     .9509327    .9860246
            polityl |   .9319431   .0287559    -2.28   0.022     .8772531    .9900427
           polityl2 |   .9838383   .0056203    -2.85   0.004     .9728842    .9949157
yrsincleaderinpower |   1.035554   .0126602     2.86   0.004     1.011035    1.060667
        v2x_execorr |   5.249923   4.154952     2.10   0.036     1.112975    24.76397
             lnoill |   .8776276   .0262353    -4.37   0.000     .8276848     .930584
        postcoldwar |   9.787698   7.173675     3.11   0.002      2.32708    41.16705
              time1 |   1.183889   .1884751     1.06   0.289     .8665623    1.617418
             timesq |   .9975108   .0022579    -1.10   0.271     .9930953    1.001946
            timecub |   1.000011   9.98e-06     1.08   0.281     .9999912     1.00003
              _cons |   2.23e-07   8.88e-07    -3.84   0.000     9.00e-11    .0005509
-------------------------------------------------------------------------------------

. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar time1 timesq timecub if  indstate==1, pa corr(ar1) eform nolog vce(robust)

note:  observations not equally spaced
       modal spacing is delta year = 1 unit
       14 groups omitted from estimation

GEE population-averaged model                   Number of obs     =      9,258
Group and time vars:          cowcode year      Number of groups  =        143
Link:                              cloglog      Obs per group:
Family:                           binomial                    min =         21
Correlation:                         AR(1)                    avg =       64.7
                                                              max =        114
                                                Wald chi2(12)     =     105.42
Scale parameter:                         1      Prob > chi2       =     0.0000

                                       (Std. Err. adjusted for clustering on cowcode)
-------------------------------------------------------------------------------------
                    |             Semirobust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.592512   .1769869     4.19   0.000     1.280805     1.98008
           gdppcthl |   1.729006   .2579446     3.67   0.000      1.29065    2.316245
          gdppcthl2 |   .9687264   .0106991    -2.88   0.004     .9479818    .9899249
            polityl |   .9387431   .0326896    -1.82   0.069     .8768101    1.005051
           polityl2 |   .9832496   .0057472    -2.89   0.004     .9720496    .9945787
yrsincleaderinpower |   1.032845   .0139451     2.39   0.017     1.005872    1.060542
        v2x_execorr |   9.072505   7.859185     2.55   0.011      1.66097    49.55558
             lnoill |   .8875043   .0294681    -3.59   0.000     .8315872    .9471814
        postcoldwar |   23.97104   22.26209     3.42   0.001     3.883017    147.9805
              time1 |    5.00036   2.395371     3.36   0.001     1.955429    12.78676
             timesq |   .9796681   .0060442    -3.33   0.001     .9678931    .9915864
            timecub |   1.000083   .0000252     3.27   0.001     1.000033    1.000132
              _cons |   1.49e-23   1.79e-22    -4.37   0.000     8.73e-34    2.54e-13
-------------------------------------------------------------------------------------

. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar time1 timesq timecub if  indstate==1, pa corr(ind) eform nolog vce(robust)

GEE population-averaged model                   Number of obs     =     10,516
Group variable:                    cowcode      Number of groups  =        157
Link:                              cloglog      Obs per group:
Family:                           binomial                    min =         21
Correlation:                   independent                    avg =       67.0
                                                              max =        114
                                                Wald chi2(12)     =     107.17
Scale parameter:                         1      Prob > chi2       =     0.0000

Pearson chi2(10516):              21868.40      Deviance          =     525.74
Dispersion (Pearson):             2.079536      Dispersion        =   .0499947

                                       (Std. Err. adjusted for clustering on cowcode)
-------------------------------------------------------------------------------------
                    |             Semirobust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.539534   .1586379     4.19   0.000     1.257995     1.88408
           gdppcthl |   1.806283   .2314335     4.61   0.000     1.405154    2.321922
          gdppcthl2 |   .9681599   .0090524    -3.46   0.001      .950579    .9860659
            polityl |   .9310612   .0288044    -2.31   0.021     .8762832    .9892634
           polityl2 |   .9837383   .0056208    -2.87   0.004     .9727833    .9948168
yrsincleaderinpower |   1.035713   .0125962     2.89   0.004     1.011317    1.060698
        v2x_execorr |   5.233018   4.148334     2.09   0.037     1.106586    24.74682
             lnoill |    .877631   .0261404    -4.38   0.000     .8278636    .9303903
        postcoldwar |     9.7927   7.158151     3.12   0.002     2.337229     41.0302
              time1 |   1.185286   .1906103     1.06   0.290     .8648451    1.624457
             timesq |   .9974958   .0022789    -1.10   0.272     .9930392    1.001972
            timecub |   1.000011   .0000101     1.08   0.282     .9999911    1.000031
              _cons |   2.15e-07   8.62e-07    -3.83   0.000     8.31e-11    .0005565
-------------------------------------------------------------------------------------

. *       RESULT:  all findings remain unchanged
. * QIC (quasilikelihood under the independence model criterion, QIC) test for model selection for a population-av
> eraged model (Model 4)
. *       For testing which correlation structure is best
. *       The correlation structure with the smallest QIC is the preferred correlation structure
. *       See Cui, James. 2007. "QIC Program and Model Selection in GEE Analyses," The Stata Journal 7, 2: 209-220
> .
. * Generate common sample
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar time1 timesq timecub if indstate==1, pa corr(ar1) vce(robust) eform nolog

note:  observations not equally spaced
       modal spacing is delta year = 1 unit
       14 groups omitted from estimation

GEE population-averaged model                   Number of obs     =      9,258
Group and time vars:          cowcode year      Number of groups  =        143
Link:                              cloglog      Obs per group:
Family:                           binomial                    min =         21
Correlation:                         AR(1)                    avg =       64.7
                                                              max =        114
                                                Wald chi2(12)     =     105.42
Scale parameter:                         1      Prob > chi2       =     0.0000

                                       (Std. Err. adjusted for clustering on cowcode)
-------------------------------------------------------------------------------------
                    |             Semirobust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.592512   .1769869     4.19   0.000     1.280805     1.98008
           gdppcthl |   1.729006   .2579446     3.67   0.000      1.29065    2.316245
          gdppcthl2 |   .9687264   .0106991    -2.88   0.004     .9479818    .9899249
            polityl |   .9387431   .0326896    -1.82   0.069     .8768101    1.005051
           polityl2 |   .9832496   .0057472    -2.89   0.004     .9720496    .9945787
yrsincleaderinpower |   1.032845   .0139451     2.39   0.017     1.005872    1.060542
        v2x_execorr |   9.072505   7.859185     2.55   0.011      1.66097    49.55558
             lnoill |   .8875043   .0294681    -3.59   0.000     .8315872    .9471814
        postcoldwar |   23.97104   22.26209     3.42   0.001     3.883017    147.9805
              time1 |    5.00036   2.395371     3.36   0.001     1.955429    12.78676
             timesq |   .9796681   .0060442    -3.33   0.001     .9678931    .9915864
            timecub |   1.000083   .0000252     3.27   0.001     1.000033    1.000132
              _cons |   1.49e-23   1.79e-22    -4.37   0.000     8.73e-34    2.54e-13
-------------------------------------------------------------------------------------

. generate sample=e(sample)

. * Testing ar1 on common sample
. qic urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldwar t
> ime1 timesq timecub if indstate==1 & sample,  corr(ar1) link(cloglog) family(binomial) robust eform nolog

GEE population-averaged model                   Number of obs     =      9,258
Group variable:                    cowcode      Number of groups  =        143
Link:                              cloglog      Obs per group:
Family:                           binomial                    min =         21
Correlation:                   independent                    avg =       64.7
                                                              max =        114
                                                Wald chi2(12)     =      69.22
Scale parameter:                         1      Prob > chi2       =     0.0000

Pearson chi2(9258):                7253.82      Deviance          =     454.23
Dispersion (Pearson):              .783519      Dispersion        =   .0490635

-------------------------------------------------------------------------------------
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.593217   .1955276     3.80   0.000     1.252596    2.026463
           gdppcthl |   1.729357   .2950501     3.21   0.001     1.237824    2.416074
          gdppcthl2 |   .9687331    .013763    -2.24   0.025     .9421302    .9960873
            polityl |   .9388087    .031726    -1.87   0.062     .8786415    1.003096
           polityl2 |   .9832393   .0061029    -2.72   0.006     .9713504    .9952738
yrsincleaderinpower |   1.033155   .0159068     2.12   0.034     1.002444    1.064807
        v2x_execorr |   9.090422   7.031894     2.85   0.004     1.995893     41.4029
             lnoill |     .88745   .0355946    -2.98   0.003     .8203576    .9600295
        postcoldwar |   23.82614   24.82182     3.04   0.002     3.092295    183.5804
              time1 |   4.980471   4.179081     1.91   0.056     .9616847    25.79337
             timesq |   .9797174   .0100253    -2.00   0.045      .960264    .9995649
            timecub |   1.000082   .0000402     2.05   0.040     1.000004    1.000161
              _cons |   1.64e-23   3.66e-22    -2.35   0.019     1.55e-42    .0001734
-------------------------------------------------------------------------------------

GEE population-averaged model                   Number of obs     =      9,258
Group and time vars:          cowcode year      Number of groups  =        143
Link:                              cloglog      Obs per group:
Family:                           binomial                    min =         21
Correlation:                         AR(1)                    avg =       64.7
                                                              max =        114
                                                Wald chi2(12)     =     105.42
Scale parameter:                         1      Prob > chi2       =     0.0000

                                       (Std. Err. adjusted for clustering on cowcode)
-------------------------------------------------------------------------------------
                    |             Semirobust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.592512   .1769869     4.19   0.000     1.280805     1.98008
           gdppcthl |   1.729006   .2579446     3.67   0.000      1.29065    2.316245
          gdppcthl2 |   .9687264   .0106991    -2.88   0.004     .9479818    .9899249
            polityl |   .9387431   .0326896    -1.82   0.069     .8768101    1.005051
           polityl2 |   .9832496   .0057472    -2.89   0.004     .9720496    .9945787
yrsincleaderinpower |   1.032845   .0139451     2.39   0.017     1.005872    1.060542
        v2x_execorr |   9.072505   7.859185     2.55   0.011      1.66097    49.55558
             lnoill |   .8875043   .0294681    -3.59   0.000     .8315872    .9471814
        postcoldwar |   23.97104   22.26209     3.42   0.001     3.883017    147.9805
              time1 |    5.00036   2.395371     3.36   0.001     1.955429    12.78676
             timesq |   .9796681   .0060442    -3.33   0.001     .9678931    .9915864
            timecub |   1.000083   .0000252     3.27   0.001     1.000033    1.000132
              _cons |   1.49e-23   1.79e-22    -4.37   0.000     8.73e-34    2.54e-13
-------------------------------------------------------------------------------------


              QIC and QIC_u
___________________________________________
Corr =                  ar1
Family =           binomial
Link =              cloglog
p =                      13
Trace =              10.712
QIC =               571.407
QIC_u =             575.984
___________________________________________

. * Testing exc on common sample)
. qic urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldwar t
> ime1 timesq timecub if indstate==1 & sample,  corr(exc) link(cloglog) family(binomial) robust eform nolog

GEE population-averaged model                   Number of obs     =      9,258
Group variable:                    cowcode      Number of groups  =        143
Link:                              cloglog      Obs per group:
Family:                           binomial                    min =         21
Correlation:                   independent                    avg =       64.7
                                                              max =        114
                                                Wald chi2(12)     =      69.22
Scale parameter:                         1      Prob > chi2       =     0.0000

Pearson chi2(9258):                7253.82      Deviance          =     454.23
Dispersion (Pearson):              .783519      Dispersion        =   .0490635

-------------------------------------------------------------------------------------
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.593217   .1955276     3.80   0.000     1.252596    2.026463
           gdppcthl |   1.729357   .2950501     3.21   0.001     1.237824    2.416074
          gdppcthl2 |   .9687331    .013763    -2.24   0.025     .9421302    .9960873
            polityl |   .9388087    .031726    -1.87   0.062     .8786415    1.003096
           polityl2 |   .9832393   .0061029    -2.72   0.006     .9713504    .9952738
yrsincleaderinpower |   1.033155   .0159068     2.12   0.034     1.002444    1.064807
        v2x_execorr |   9.090422   7.031894     2.85   0.004     1.995893     41.4029
             lnoill |     .88745   .0355946    -2.98   0.003     .8203576    .9600295
        postcoldwar |   23.82614   24.82182     3.04   0.002     3.092295    183.5804
              time1 |   4.980471   4.179081     1.91   0.056     .9616847    25.79337
             timesq |   .9797174   .0100253    -2.00   0.045      .960264    .9995649
            timecub |   1.000082   .0000402     2.05   0.040     1.000004    1.000161
              _cons |   1.64e-23   3.66e-22    -2.35   0.019     1.55e-42    .0001734
-------------------------------------------------------------------------------------

GEE population-averaged model                   Number of obs     =      9,258
Group variable:                    cowcode      Number of groups  =        143
Link:                              cloglog      Obs per group:
Family:                           binomial                    min =         21
Correlation:                  exchangeable                    avg =       64.7
                                                              max =        114
                                                Wald chi2(12)     =     104.92
Scale parameter:                         1      Prob > chi2       =     0.0000

                                       (Std. Err. adjusted for clustering on cowcode)
-------------------------------------------------------------------------------------
                    |             Semirobust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.594632   .1774098     4.19   0.000     1.282214    1.983172
           gdppcthl |     1.7263   .2577657     3.66   0.000     1.288301     2.31321
          gdppcthl2 |   .9688383   .0106994    -2.87   0.004     .9480931    .9900374
            polityl |   .9395631   .0326142    -1.80   0.073     .8777665     1.00571
           polityl2 |    .983273   .0057425    -2.89   0.004     .9720821    .9945927
yrsincleaderinpower |   1.032873   .0140321     2.38   0.017     1.005733    1.060745
        v2x_execorr |   9.116573    7.88956     2.55   0.011     1.671839    49.71285
             lnoill |   .8875373   .0296219    -3.57   0.000     .8313376    .9475361
        postcoldwar |   23.94649   22.29661     3.41   0.001     3.860885    148.5241
              time1 |   4.999492   2.394948     3.36   0.001     1.955095    12.78451
             timesq |   .9796692   .0060469    -3.33   0.001     .9678889    .9915929
            timecub |   1.000083   .0000253     3.27   0.001     1.000033    1.000132
              _cons |   1.48e-23   1.78e-22    -4.38   0.000     8.88e-34    2.48e-13
-------------------------------------------------------------------------------------


              QIC and QIC_u
___________________________________________
Corr =                  exc
Family =           binomial
Link =              cloglog
p =                      13
Trace =              10.717
QIC =               571.444
QIC_u =             576.009
___________________________________________

. * Testing ind on common sample
. qic urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldwar t
> ime1 timesq timecub if indstate==1 & sample,  corr(ind) link(cloglog) family(binomial) robust eform nolog

GEE population-averaged model                   Number of obs     =      9,258
Group variable:                    cowcode      Number of groups  =        143
Link:                              cloglog      Obs per group:
Family:                           binomial                    min =         21
Correlation:                   independent                    avg =       64.7
                                                              max =        114
                                                Wald chi2(12)     =      69.22
Scale parameter:                         1      Prob > chi2       =     0.0000

Pearson chi2(9258):                7253.82      Deviance          =     454.23
Dispersion (Pearson):              .783519      Dispersion        =   .0490635

-------------------------------------------------------------------------------------
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.593217   .1955276     3.80   0.000     1.252596    2.026463
           gdppcthl |   1.729357   .2950501     3.21   0.001     1.237824    2.416074
          gdppcthl2 |   .9687331    .013763    -2.24   0.025     .9421302    .9960873
            polityl |   .9388087    .031726    -1.87   0.062     .8786415    1.003096
           polityl2 |   .9832393   .0061029    -2.72   0.006     .9713504    .9952738
yrsincleaderinpower |   1.033155   .0159068     2.12   0.034     1.002444    1.064807
        v2x_execorr |   9.090422   7.031894     2.85   0.004     1.995893     41.4029
             lnoill |     .88745   .0355946    -2.98   0.003     .8203576    .9600295
        postcoldwar |   23.82614   24.82182     3.04   0.002     3.092295    183.5804
              time1 |   4.980471   4.179081     1.91   0.056     .9616847    25.79337
             timesq |   .9797174   .0100253    -2.00   0.045      .960264    .9995649
            timecub |   1.000082   .0000402     2.05   0.040     1.000004    1.000161
              _cons |   1.64e-23   3.66e-22    -2.35   0.019     1.55e-42    .0001734
-------------------------------------------------------------------------------------

GEE population-averaged model                   Number of obs     =      9,258
Group variable:                    cowcode      Number of groups  =        143
Link:                              cloglog      Obs per group:
Family:                           binomial                    min =         21
Correlation:                   independent                    avg =       64.7
                                                              max =        114
                                                Wald chi2(12)     =     105.40
Scale parameter:                         1      Prob > chi2       =     0.0000

Pearson chi2(9258):                7253.82      Deviance          =     454.23
Dispersion (Pearson):              .783519      Dispersion        =   .0490635

                                       (Std. Err. adjusted for clustering on cowcode)
-------------------------------------------------------------------------------------
                    |             Semirobust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.593217   .1771497     4.19   0.000     1.281238    1.981162
           gdppcthl |   1.729357   .2580324     3.67   0.000      1.29086    2.316807
          gdppcthl2 |   .9687331   .0106972    -2.88   0.004     .9479922    .9899278
            polityl |   .9388087   .0327724    -1.81   0.070      .876724     1.00529
           polityl2 |   .9832393   .0057572    -2.89   0.004       .97202    .9945882
yrsincleaderinpower |   1.033155   .0139624     2.41   0.016     1.006148    1.060886
        v2x_execorr |   9.090422    7.87991     2.55   0.011     1.662385    49.70918
             lnoill |     .88745   .0294754    -3.60   0.000     .8315196    .9471425
        postcoldwar |   23.82614   22.09665     3.42   0.001     3.869357     146.713
              time1 |   4.980471   2.384467     3.35   0.001     1.948706      12.729
             timesq |   .9797174   .0060423    -3.32   0.001     .9679459     .991632
            timecub |   1.000082   .0000252     3.26   0.001     1.000033    1.000132
              _cons |   1.64e-23   1.97e-22    -4.37   0.000     9.81e-34    2.74e-13
-------------------------------------------------------------------------------------


              QIC and QIC_u
___________________________________________
Corr =                  ind
Family =           binomial
Link =              cloglog
p =                      13
Trace =              10.722
QIC =               571.341
QIC_u =             575.897
___________________________________________

. * an exchangeable (equal correlation) structure proved the most efficient (lowest qic)
. drop sample

. 
. * +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
. * Testing robustness of specification to inclusion or exclusion of variables
. * +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
. * checkrob procedure using Model 4
. * Testing if signs are stable and z-statistics for Beta/S.E. of Beta are >=1.96 (i.e., .05 level of significance
> )
. *       or >=1.65 (i.e., .10 levels of significance)
. *   z-statistics were calculated in Excel (as comma-separated tables) from the output produced by checkrob 
. * BE AWARE that running checkrob can take a considerable amount of time
. * There is also sometimes a glitch in the checkrob procedure
. *    In the tables produced, the results for the first variable of a quadratic specification sometimes falsely 
. *       include some results from the squared variable    
. *    MUST VISUALLY INSPECT AND, IF PRESENT, HAND-CORRECT AND RECALCULATE RESULTS FOR THOSE VARIABLES
. *    (This was done for gdppcthl and gdppcthl2 and polityl and polityl2 below)
. * +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
. ** Model 4 (testing additional elements)
. * postcoldwar kept in all specifications as time control
. * I HAVE PROVIDED THE EXCEL TABLE WITH RESULTS OF THE TEST BELOW, CORRECTED FOR THE ABOVE PROBLEMS
. * THE RESULTS ARE PRESENTED WITH OTHER ROBUSTNESS TESTS FILES IN THE FILE checkrob.results.chapter3.xlsx
. * COMMAND USED: checkrob 1 8 ch3tab1mod4.txt: xtcloglog urbancivicny postcoldwar lnpopl gdppcthl gdppcthl2 polit
> yl polityl2 yrsincleaderinpower v2x_execorr lnoill if indstate==1
. * RESULTS:  All variables had stable signs and were statistically significant, with exceptions of v2x_execorr an
> d lnoill
. *               v2x_execorr:  no change of signs, but significant at the .05 level in only 75 percent of specifi
> cations (problematic when gdppcthl is dropped)
. *               lnoill:  change of signs in 12.5% of specifications (when gdppcthl is not included)
. *                                significant at the .05 or .10 levels only in 40.5% of specifications (insignifi
> cant when either gdppcthl or polityl are dropped)
. 
. * Follow-up test: Likelihood ratio test for whether inclusion of oil production improves the accuracy of the mod
> el
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar if indstate==1, eform nolog

Random-effects complementary log-log model      Number of obs     =     10,516
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       67.0
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(9)      =      82.87
Log likelihood  = -263.63109                    Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.518909    .165332     3.84   0.000     1.227097    1.880115
           gdppcthl |   1.824338   .2727876     4.02   0.000     1.360905    2.445585
          gdppcthl2 |   .9673954   .0114607    -2.80   0.005     .9451915    .9901208
            polityl |   .9269684   .0274565    -2.56   0.010     .8746868    .9823749
           polityl2 |   .9841294   .0055479    -2.84   0.005     .9733155    .9950634
yrsincleaderinpower |   1.033032   .0141883     2.37   0.018     1.005595    1.061219
        v2x_execorr |   5.163556   3.462523     2.45   0.014     1.387264    19.21935
             lnoill |   .8778001   .0318752    -3.59   0.000     .8174972    .9425512
        postcoldwar |   6.900293   2.971396     4.49   0.000     2.967046    16.04763
              _cons |   5.73e-06   7.31e-06    -9.46   0.000     4.70e-07    .0000698
--------------------+----------------------------------------------------------------
           /lnsig2u |  -10.47004   26.33272                     -62.08123    41.14115
--------------------+----------------------------------------------------------------
            sigma_u |   .0053267   .0701335                      3.31e-14    8.58e+08
                rho |   .0000172   .0004542                      6.64e-28           1
-------------------------------------------------------------------------------------
LR test of rho=0: chibar2(01) = 5.6e-05                Prob >= chibar2 = 0.497

. generate sample=e(sample)

. estimates store mod1

. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr  postcoldwar i
> f indstate==1 & sample==1, eform nolog

Random-effects complementary log-log model      Number of obs     =     10,516
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       67.0
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(8)      =      59.67
Log likelihood  = -269.72432                    Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.308963   .1358553     2.59   0.009     1.068028    1.604251
           gdppcthl |   1.505077   .2092764     2.94   0.003     1.146045    1.976586
          gdppcthl2 |    .976202   .0106457    -2.21   0.027     .9555581    .9972918
            polityl |   .9419149   .0292063    -1.93   0.054     .8863764    1.000933
           polityl2 |   .9849077   .0057389    -2.61   0.009     .9737235    .9962203
yrsincleaderinpower |   1.034489   .0158226     2.22   0.027     1.003937     1.06597
        v2x_execorr |    3.16023   2.197507     1.65   0.098     .8087691    12.34846
        postcoldwar |   6.348414   2.798882     4.19   0.000     2.675398    15.06406
              _cons |   .0000287   .0000365    -8.21   0.000     2.36e-06    .0003481
--------------------+----------------------------------------------------------------
           /lnsig2u |  -1.559771   2.373613                     -6.211966    3.092425
--------------------+----------------------------------------------------------------
            sigma_u |   .4584585   .5441015                      .0447805    4.693658
                rho |   .1132996   .2384599                      .0012176    .9305213
-------------------------------------------------------------------------------------
LR test of rho=0: chibar2(01) = 0.19                   Prob >= chibar2 = 0.329

. estimates store mod2

. lrtest mod1 mod2

Likelihood-ratio test                                 LR chi2(1)  =     12.19
(Assumption: mod2 nested in mod1)                     Prob > chi2 =    0.0005

. drop _est_mod1 _est_mod2

. * RESULT:  Significant improvement of the model when lnoill is included
. 
. * ++++++++++++++++++++++++++++++++++++++++++++
. * Tests for omitted variable bias in Model 4
. * ++++++++++++++++++++++++++++++++++++++++++++
. * Logged military spending per soldier: lnmilexppersoldthl
. * Bivariate, controlling for time
. xtcloglog urbancivicny lnmilexppersoldthl time1 timesq timecub if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =      9,925
Group variable: cowcode                         Number of groups  =        163

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          3
                                                              avg =       60.9
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      27.14
Log pseudolikelihood  = -282.09125              Prob > chi2       =     0.0000

                                    (Std. Err. adjusted for 163 clusters in cowcode)
------------------------------------------------------------------------------------
                   |               Robust
      urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------------+----------------------------------------------------------------
lnmilexppersoldthl |   .7879725   .0844068    -2.22   0.026     .6387505     .972055
             time1 |   .9620568   .0854738    -0.44   0.663     .8083061    1.145053
            timesq |   1.001306   .0014532     0.90   0.369     .9984616    1.004158
           timecub |   .9999939   7.28e-06    -0.84   0.400     .9999796    1.000008
             _cons |    .000701   .0013567    -3.75   0.000     .0000158    .0311196
-------------------+----------------------------------------------------------------
          /lnsig2u |  -10.19514   7250.598                     -14221.11    14200.72
-------------------+----------------------------------------------------------------
           sigma_u |   .0061116   22.15631                             0           .
               rho |   .0000227   .1646314                             0           .
------------------------------------------------------------------------------------

. * Result:  negative and statistically signficant at the .05 level in bivariate relationship, controlling for tim
> e
. * Multivariate cloglog panel, Model 4 from Table 3.1
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar lnmilexppersoldthl if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =      9,442
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          3
                                                              avg =       60.1
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      81.09
Log pseudolikelihood  = -245.86764              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 157 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.508538   .1503821     4.12   0.000     1.240802    1.834046
           gdppcthl |   1.877562   .2456885     4.81   0.000     1.452815    2.426489
          gdppcthl2 |    .965933   .0094272    -3.55   0.000     .9476318    .9845878
            polityl |   .9159754   .0342922    -2.34   0.019     .8511706    .9857141
           polityl2 |   .9841753   .0056707    -2.77   0.006     .9731234    .9953527
yrsincleaderinpower |   1.031159   .0136979     2.31   0.021     1.004658    1.058358
        v2x_execorr |   5.427939   5.094584     1.80   0.072     .8624015    34.16334
             lnoill |   .8804331   .0284401    -3.94   0.000     .8264194    .9379771
        postcoldwar |   6.541995    3.19343     3.85   0.000     2.513055    17.03015
 lnmilexppersoldthl |   .9811839   .1686608    -0.11   0.912       .70054    1.374257
              _cons |   5.95e-06   8.59e-06    -8.34   0.000     3.52e-07    .0001006
--------------------+----------------------------------------------------------------
           /lnsig2u |  -11.06566    35909.4                      -70392.2    70370.06
--------------------+----------------------------------------------------------------
            sigma_u |   .0039548   71.00708                             0           .
                rho |   9.51e-06   .3414275                             0           .
-------------------------------------------------------------------------------------

. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -245.86764     -245.86761     -245.86761
likelihood                   .00003277      .00003277   Difference
                            -1.333e-07     -1.333e-07   Relative difference
-----------------------------------------------------
urbancivicny: .41114122      .41113425      .41113425
  lnpopl                    -6.967e-06     -6.967e-06   Difference
                            -.00001695     -.00001695   Relative difference
-----------------------------------------------------
urbancivicny: .62997439      .62999964      .62999964
gdppcthl                     .00002525      .00002525   Difference
                             .00004009      .00004009   Relative difference
-----------------------------------------------------
urbancivicny:-.03466075     -.03466317     -.03466317
gdppcthl2                   -2.420e-06     -2.420e-06   Difference
                             .00006981      .00006981   Relative difference
-----------------------------------------------------
urbancivicny: -.0877658      -.0877805      -.0877805
 polityl                    -.00001469     -.00001469   Difference
                             .00016743      .00016743   Relative difference
-----------------------------------------------------
urbancivicny:-.01595124       -.015951       -.015951
polityl2                     2.363e-07      2.363e-07   Difference
                            -.00001481     -.00001481   Relative difference
-----------------------------------------------------
urbancivicny: .03068297      .03069429      .03069429
yrsinclead~r                 .00001132      .00001132   Difference
                             .00036898      .00036898   Relative difference
-----------------------------------------------------
urbancivicny: 1.6915594      1.6912104      1.6912104
v2x_execorr                 -.00034896     -.00034896   Difference
                             -.0002063      -.0002063   Relative difference
-----------------------------------------------------
urbancivicny:-.12734132     -.12734203     -.12734203
  lnoill                    -7.064e-07     -7.064e-07   Difference
                             5.547e-06      5.547e-06   Relative difference
-----------------------------------------------------
urbancivicny: 1.8782422      1.8784157      1.8784157
postcoldwar                  .00017348      .00017348   Difference
                             .00009236      .00009236   Relative difference
-----------------------------------------------------
urbancivicny:-.01899542     -.01903773     -.01903773
lnmilexppe~l                -.00004231     -.00004231   Difference
                             .00222715      .00222715   Relative difference
-----------------------------------------------------
urbancivicny:-12.031929      -12.03227      -12.03227
   _cons                    -.00034099     -.00034099   Difference
                             .00002834      .00002834   Relative difference
-----------------------------------------------------
lnsig2u:     -11.065655     -12.935561     -12.935561
   _cons                    -1.8699056     -1.8699056   Difference
                             .16898282      .16898282   Relative difference
-----------------------------------------------------

. * Failed quadchk--recalculate using pooled model
. cloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldw
> ar lnmilexppersoldthl if indstate==1, vce(robust) eform nolog

Complementary log-log regression                Number of obs     =      9,442
                                                Zero outcomes     =      9,395
                                                Nonzero outcomes  =         47

                                                Wald chi2(10)     =      69.18
Log pseudolikelihood =  -245.8676               Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.508533   .1654569     3.75   0.000     1.216731    1.870316
           gdppcthl |   1.877587   .2795405     4.23   0.000     1.402396    2.513792
          gdppcthl2 |   .9659318   .0100881    -3.32   0.001     .9463604     .985908
            polityl |   .9159682   .0313853    -2.56   0.010     .8564743    .9795948
           polityl2 |   .9841754   .0056009    -2.80   0.005     .9732589    .9952145
yrsincleaderinpower |   1.031165   .0125572     2.52   0.012     1.006845    1.056072
        v2x_execorr |   5.426927   4.488472     2.05   0.041      1.07287    27.45118
             lnoill |   .8804328   .0307872    -3.64   0.000     .8221123    .9428906
        postcoldwar |     6.5426   2.955942     4.16   0.000     2.698846    15.86071
 lnmilexppersoldthl |   .9811617   .1635099    -0.11   0.909      .707764    1.360169
              _cons |   5.95e-06   8.63e-06    -8.29   0.000     3.47e-07    .0001022
-------------------------------------------------------------------------------------

. *  Result:  
. *    --lnmilexppersoldthl statistically insignificant
. *    --no sign changes or change in statistical significance of all other variables
. 
. * Military soldiers per 1000 pop:  milperthousl
. * Bivariate, controlling for time
. xtcloglog urbancivicny milperthousl time1 timesq timecub if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,038
Group variable: cowcode                         Number of groups  =        163

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       67.7
                                                              max =        115

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      22.93
Log pseudolikelihood  = -300.94917              Prob > chi2       =     0.0001

                              (Std. Err. adjusted for 163 clusters in cowcode)
------------------------------------------------------------------------------
             |               Robust
urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
milperthousl |   1.006657   .0061963     1.08   0.281     .9945852    1.018875
       time1 |   .9527984   .0796308    -0.58   0.563     .8088375    1.122382
      timesq |   1.001456   .0013763     1.06   0.290     .9987619    1.004157
     timecub |    .999993   6.86e-06    -1.02   0.308     .9999796    1.000006
       _cons |    .000485   .0008798    -4.21   0.000     .0000138    .0169824
-------------+----------------------------------------------------------------
    /lnsig2u |   -1.57399   1.210761                     -3.947038    .7990586
-------------+----------------------------------------------------------------
     sigma_u |   .4552107   .2755757                      .1389669    1.491123
         rho |    .111879   .1203038                      .0116039    .5747745
------------------------------------------------------------------------------

. * Result:  postive and statistically insignificant in bivariate relationship, controlling for time
. * Multivariate cloglog panel, Model 4 from Table 3.1
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar milperthousl if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,174
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       64.8
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =     100.76
Log pseudolikelihood  = -262.00125              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 157 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |    1.49426   .1432002     4.19   0.000     1.238376    1.803017
           gdppcthl |   1.956732   .2621333     5.01   0.000     1.504875    2.544266
          gdppcthl2 |   .9631455   .0093453    -3.87   0.000     .9450021    .9816372
            polityl |    .919448   .0300766    -2.57   0.010     .8623489    .9803279
           polityl2 |   .9850346   .0054946    -2.70   0.007     .9743241    .9958629
yrsincleaderinpower |   1.036047   .0129243     2.84   0.005     1.011023     1.06169
        v2x_execorr |   4.981983   4.359783     1.84   0.067     .8964115    27.68835
             lnoill |   .8682425   .0262672    -4.67   0.000     .8182564    .9212821
        postcoldwar |   6.885376   2.985909     4.45   0.000     2.943039    16.10866
       milperthousl |   .9769901   .0161216    -1.41   0.158     .9458979    1.009104
              _cons |   6.95e-06   9.21e-06    -8.96   0.000     5.16e-07    .0000935
--------------------+----------------------------------------------------------------
           /lnsig2u |  -11.08886   27008.58                     -52946.94    52924.76
--------------------+----------------------------------------------------------------
            sigma_u |   .0039092    52.7905                             0           .
                rho |   9.29e-06   .2509072                             0           .
-------------------------------------------------------------------------------------

. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -262.00125     -262.00122     -262.00122
likelihood                   .00002423      .00002423   Difference
                            -9.249e-08     -9.249e-08   Relative difference
-----------------------------------------------------
urbancivicny: .40163115      .40161671      .40161671
  lnpopl                    -.00001444     -.00001444   Difference
                            -.00003595     -.00003595   Relative difference
-----------------------------------------------------
urbancivicny: .67127585      .67127814      .67127814
gdppcthl                     2.289e-06      2.289e-06   Difference
                             3.410e-06      3.410e-06   Relative difference
-----------------------------------------------------
urbancivicny:-.03755078     -.03755188     -.03755188
gdppcthl2                   -1.104e-06     -1.104e-06   Difference
                             .00002939      .00002939   Relative difference
-----------------------------------------------------
urbancivicny:-.08398174     -.08399469     -.08399469
 polityl                    -.00001295     -.00001295   Difference
                             .00015421      .00015421   Relative difference
-----------------------------------------------------
urbancivicny:-.01507848     -.01507923     -.01507923
polityl2                    -7.567e-07     -7.567e-07   Difference
                             .00005019      .00005019   Relative difference
-----------------------------------------------------
urbancivicny: .03541266      .03542271      .03542271
yrsinclead~r                 .00001005      .00001005   Difference
                             .00028377      .00028377   Relative difference
-----------------------------------------------------
urbancivicny:  1.605828      1.6054304      1.6054304
v2x_execorr                 -.00039761     -.00039761   Difference
                             -.0002476      -.0002476   Relative difference
-----------------------------------------------------
urbancivicny:-.14128425     -.14127878     -.14127878
  lnoill                     5.473e-06      5.473e-06   Difference
                            -.00003874     -.00003874   Relative difference
-----------------------------------------------------
urbancivicny: 1.9293997       1.929512       1.929512
postcoldwar                  .00011221      .00011221   Difference
                             .00005816      .00005816   Relative difference
-----------------------------------------------------
urbancivicny:-.02327872      -.0232813      -.0232813
milperthousl                -2.579e-06     -2.579e-06   Difference
                             .00011078      .00011078   Relative difference
-----------------------------------------------------
urbancivicny:-11.877097     -11.877254     -11.877254
   _cons                    -.00015681     -.00015681   Difference
                              .0000132       .0000132   Relative difference
-----------------------------------------------------
lnsig2u:     -11.088863     -12.955897     -12.955897
   _cons                    -1.8670337     -1.8670338   Difference
                             .16837017      .16837017   Relative difference
-----------------------------------------------------

. * Failed quadchk--recalculate using pooled model
. cloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldw
> ar milperthousl if indstate==1, vce(robust) eform nolog

Complementary log-log regression                Number of obs     =     10,174
                                                Zero outcomes     =     10,124
                                                Nonzero outcomes  =         50

                                                Wald chi2(10)     =      73.56
Log pseudolikelihood = -262.00122               Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.494249   .1580583     3.80   0.000     1.214464    1.838489
           gdppcthl |   1.956735   .2834476     4.63   0.000     1.473089    2.599171
          gdppcthl2 |   .9631449   .0099531    -3.63   0.000     .9438335    .9828515
            polityl |   .9194417    .029975    -2.58   0.010     .8625295    .9801092
           polityl2 |   .9850342   .0054106    -2.75   0.006     .9744865    .9956961
yrsincleaderinpower |   1.036053   .0131453     2.79   0.005     1.010606     1.06214
        v2x_execorr |   4.980927   3.880098     2.06   0.039     1.082011    22.92919
             lnoill |    .868245   .0310405    -3.95   0.000     .8094893    .9312654
        postcoldwar |   6.885788   2.895671     4.59   0.000     3.019945    15.70031
       milperthousl |   .9769888   .0147518    -1.54   0.123     .9484995    1.006334
              _cons |   6.95e-06   9.81e-06    -8.41   0.000     4.37e-07    .0001105
-------------------------------------------------------------------------------------

. *  Result  
. *    --milperthousl marginally statistically significant at the .10 level
. *    --no sign change or change in statistical significance of all other variables
. 
. * Mean years of schooling: totalyrsschooll
. * Bivariate, controlling for time
. xtcloglog urbancivicny totalyrsschooll time1 timesq timecub if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =      7,132
Group variable: cowcode                         Number of groups  =        136

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         19
                                                              avg =       52.4
                                                              max =         61

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      14.52
Log pseudolikelihood  = -228.32667              Prob > chi2       =     0.0058

                                 (Std. Err. adjusted for 136 clusters in cowcode)
---------------------------------------------------------------------------------
                |               Robust
   urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
totalyrsschooll |     .91524   .0470484    -1.72   0.085     .8275201    1.012258
          time1 |   1.604486   1.496109     0.51   0.612     .2580058    9.977971
         timesq |   .9949528   .0109657    -0.46   0.646     .9736909    1.016679
        timecub |    1.00002    .000043     0.46   0.647     .9999354    1.000104
          _cons |   1.00e-09   2.62e-08    -0.79   0.427     6.07e-32    1.66e+13
----------------+----------------------------------------------------------------
       /lnsig2u |  -2.867047    5.30092                     -13.25666    7.522566
----------------+----------------------------------------------------------------
        sigma_u |   .2384672   .6320478                      .0013224    43.00356
            rho |   .0334155   .1712141                      1.06e-06    .9991113
---------------------------------------------------------------------------------

. * Result:  negative and statistically insignificant in bivariate relationship, controlling for time
. * Multivariate cloglog panel, Model 4 from Table 3.1
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar totalyrsschooll if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =      6,853
Group variable: cowcode                         Number of groups  =        132

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         18
                                                              avg =       51.9
                                                              max =         61

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      74.55
Log pseudolikelihood  = -189.65536              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 132 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.612605   .3014676     2.56   0.011     1.117893    2.326245
           gdppcthl |   1.759588   .4210221     2.36   0.018     1.100883    2.812423
          gdppcthl2 |     .96753   .0145951    -2.19   0.029     .9393428     .996563
            polityl |   .8822173   .0459945    -2.40   0.016     .7965225    .9771317
           polityl2 |   .9823618   .0080239    -2.18   0.029     .9667605    .9982149
yrsincleaderinpower |   1.030952   .0172164     1.83   0.068     .9977546    1.065254
        v2x_execorr |   10.12854   15.04034     1.56   0.119     .5515173    186.0094
             lnoill |   .8734333   .0534986    -2.21   0.027     .7746275    .9848421
        postcoldwar |   4.350833   2.145254     2.98   0.003     1.655285    11.43595
    totalyrsschooll |   1.041538   .1156556     0.37   0.714     .8378286    1.294778
              _cons |   2.77e-06   5.58e-06    -6.34   0.000     5.29e-08    .0001445
--------------------+----------------------------------------------------------------
           /lnsig2u |  -12.19078     291613                     -571563.2    571538.8
--------------------+----------------------------------------------------------------
            sigma_u |   .0022532   328.5357                             0           .
                rho |   3.09e-06   .9000508                             0           .
-------------------------------------------------------------------------------------

. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -189.65536     -189.65535     -189.65535
likelihood                   9.978e-06      9.978e-06   Difference
                            -5.261e-08     -5.261e-08   Relative difference
-----------------------------------------------------
urbancivicny: .47785059      .47778627      .47778627
  lnpopl                    -.00006432     -.00006432   Difference
                             -.0001346      -.0001346   Relative difference
-----------------------------------------------------
urbancivicny: .56507957      .56501499      .56501499
gdppcthl                    -.00006458     -.00006458   Difference
                            -.00011429     -.00011429   Relative difference
-----------------------------------------------------
urbancivicny:-.03300887     -.03300533     -.03300533
gdppcthl2                    3.539e-06      3.539e-06   Difference
                            -.00010722     -.00010722   Relative difference
-----------------------------------------------------
urbancivicny:-.12531684     -.12532655     -.12532655
 polityl                    -9.705e-06     -9.705e-06   Difference
                             .00007744      .00007744   Relative difference
-----------------------------------------------------
urbancivicny:-.01779557     -.01779663     -.01779663
polityl2                    -1.056e-06     -1.056e-06   Difference
                             .00005936      .00005936   Relative difference
-----------------------------------------------------
urbancivicny: .03048253      .03048815      .03048815
yrsinclead~r                 5.620e-06      5.620e-06   Difference
                             .00018436      .00018436   Relative difference
-----------------------------------------------------
urbancivicny: 2.3153576      2.3149903      2.3149903
v2x_execorr                 -.00036731     -.00036731   Difference
                            -.00015864     -.00015864   Relative difference
-----------------------------------------------------
urbancivicny:-.13532351     -.13530758     -.13530758
  lnoill                     .00001593      .00001593   Difference
                            -.00011772     -.00011772   Relative difference
-----------------------------------------------------
urbancivicny: 1.4703674      1.4704218      1.4704218
postcoldwar                  .00005444      .00005444   Difference
                             .00003703      .00003703   Relative difference
-----------------------------------------------------
urbancivicny: .04069866      .04072657      .04072657
totalyrss~ll                 .00002791      .00002791   Difference
                             .00068587      .00068587   Relative difference
-----------------------------------------------------
urbancivicny:-12.798292      -12.79786      -12.79786
   _cons                     .00043166      .00043166   Difference
                            -.00003373     -.00003373   Relative difference
-----------------------------------------------------
lnsig2u:      -12.19078     -15.063088     -15.063088
   _cons                    -2.8723075     -2.8723075   Difference
                             .23561309      .23561309   Relative difference
-----------------------------------------------------

. * Failed quadchk--recalculate using pooled model
. cloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldw
> ar totalyrsschooll if indstate==1, vce(robust) eform nolog

Complementary log-log regression                Number of obs     =      6,853
                                                Zero outcomes     =      6,816
                                                Nonzero outcomes  =         37

                                                Wald chi2(10)     =      57.49
Log pseudolikelihood = -189.65535               Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.612569    .234512     3.29   0.001     1.212635    2.144403
           gdppcthl |   1.759545   .3474985     2.86   0.004     1.194796    2.591237
          gdppcthl2 |   .9675314   .0140561    -2.27   0.023     .9403703    .9954769
            polityl |   .8822144   .0386202    -2.86   0.004     .8096766    .9612507
           polityl2 |   .9823615   .0076424    -2.29   0.022     .9674964     .997455
yrsincleaderinpower |   1.030954   .0158624     1.98   0.048     1.000328    1.062517
        v2x_execorr |   10.12724   10.59462     2.21   0.027     1.303185    78.70028
             lnoill |   .8734382   .0416227    -2.84   0.005      .795553    .9589483
        postcoldwar |   4.350914   2.099248     3.05   0.002     1.690009    11.20139
    totalyrsschooll |   1.041548     .08298     0.51   0.609     .8909727    1.217572
              _cons |   2.77e-06   5.59e-06    -6.33   0.000     5.26e-08    .0001454
-------------------------------------------------------------------------------------

. *  Result  
. *    --sample is highly reduced since data for totalyrsschooll only exist since 1951 (n=6,853)
. *    --totalyrsschooll is statistically insignificant 
. *    --no sign change or change in statistical significance for all other variables
. 
. * Levels of urbanization:  percurbanl
. * Bivariate, controlling for time
. xtcloglog urbancivicny percurbanl time1 timesq timecub if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,318
Group variable: cowcode                         Number of groups  =        162

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         22
                                                              avg =       69.9
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      25.18
Log pseudolikelihood  = -306.45261              Prob > chi2       =     0.0000

                              (Std. Err. adjusted for 162 clusters in cowcode)
------------------------------------------------------------------------------
             |               Robust
urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
  percurbanl |   .9942573    .006083    -0.94   0.347      .982406    1.006252
       time1 |    .936245   .0768522    -0.80   0.422     .7971098    1.099666
      timesq |   1.001762   .0013701     1.29   0.198     .9990805    1.004451
     timecub |   .9999916   6.83e-06    -1.23   0.218     .9999782    1.000005
       _cons |   .0007806   .0013615    -4.10   0.000     .0000256    .0238283
-------------+----------------------------------------------------------------
    /lnsig2u |  -1.781988   1.474068                     -4.671108    1.107133
-------------+----------------------------------------------------------------
     sigma_u |   .4102479   .3023667                      .0967568    1.739446
         rho |   .0928192   .1241222                      .0056591    .6478115
------------------------------------------------------------------------------

. * Result:  statistically insignificant in bivariate relationship, controlling for time
. * Multivariate cloglog panel, Model 4 from Table 3.1
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar percurbanl if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,421
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       66.4
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      97.12
Log pseudolikelihood  = -263.52536              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 157 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.522103   .1513272     4.23   0.000     1.252615    1.849568
           gdppcthl |   1.804233   .2642903     4.03   0.000     1.353959    2.404249
          gdppcthl2 |    .967959   .0092175    -3.42   0.001     .9500605    .9861947
            polityl |   .9268481   .0306156    -2.30   0.021     .8687437    .9888386
           polityl2 |   .9840493   .0054686    -2.89   0.004     .9733892    .9948262
yrsincleaderinpower |   1.032788   .0126262     2.64   0.008     1.008335    1.057834
        v2x_execorr |   5.247074   4.572288     1.90   0.057     .9510035    28.95024
             lnoill |   .8773833    .025648    -4.47   0.000     .8285272    .9291203
        postcoldwar |   6.753046   2.945594     4.38   0.000      2.87221    15.87754
         percurbanl |   1.001048    .009967     0.11   0.916     .9817023    1.020775
              _cons |   5.60e-06   8.15e-06    -8.31   0.000     3.23e-07     .000097
--------------------+----------------------------------------------------------------
           /lnsig2u |  -10.19778   10891.01                     -21356.18    21335.78
--------------------+----------------------------------------------------------------
            sigma_u |   .0061035   33.23679                             0           .
                rho |   .0000226   .2466391                             0           .
-------------------------------------------------------------------------------------

. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -263.52536     -263.52533     -263.52533
likelihood                   .00003304      .00003304   Difference
                            -1.254e-07     -1.254e-07   Relative difference
-----------------------------------------------------
urbancivicny:  .4200929      .42008655      .42008655
  lnpopl                    -6.350e-06     -6.350e-06   Difference
                            -.00001512     -.00001512   Relative difference
-----------------------------------------------------
urbancivicny: .59013531      .59012003      .59012003
gdppcthl                    -.00001529     -.00001529   Difference
                             -.0000259      -.0000259   Relative difference
-----------------------------------------------------
urbancivicny:-.03256557     -.03256615     -.03256615
gdppcthl2                   -5.786e-07     -5.786e-07   Difference
                             .00001777      .00001777   Relative difference
-----------------------------------------------------
urbancivicny:-.07596564      -.0759812      -.0759812
 polityl                    -.00001556     -.00001556   Difference
                             .00020483      .00020483   Relative difference
-----------------------------------------------------
urbancivicny:-.01607929      -.0160793      -.0160793
polityl2                    -1.489e-08     -1.489e-08   Difference
                             9.259e-07      9.259e-07   Relative difference
-----------------------------------------------------
urbancivicny:  .0322619      .03227483      .03227483
yrsinclead~r                 .00001293      .00001293   Difference
                             .00040067      .00040067   Relative difference
-----------------------------------------------------
urbancivicny: 1.6576705      1.6573313      1.6573313
v2x_execorr                 -.00033918     -.00033918   Difference
                            -.00020461     -.00020461   Relative difference
-----------------------------------------------------
urbancivicny:-.13081136     -.13081278     -.13081278
  lnoill                    -1.423e-06     -1.423e-06   Difference
                             .00001088      .00001088   Relative difference
-----------------------------------------------------
urbancivicny: 1.9099936      1.9101112      1.9101112
postcoldwar                  .00011759      .00011759   Difference
                             .00006156      .00006156   Relative difference
-----------------------------------------------------
urbancivicny: .00104744      .00104966      .00104966
percurbanl                   2.225e-06      2.225e-06   Difference
                             .00212471      .00212471   Relative difference
-----------------------------------------------------
urbancivicny:-12.092939     -12.093373     -12.093373
   _cons                    -.00043398     -.00043398   Difference
                             .00003589      .00003589   Relative difference
-----------------------------------------------------
lnsig2u:     -10.197776     -12.057939     -12.057939
   _cons                    -1.8601632     -1.8601632   Difference
                             .18240871      .18240871   Relative difference
-----------------------------------------------------

. * Failed quadchk--recalculate using pooled model
. cloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldw
> ar percurbanl if indstate==1, vce(robust) eform nolog

Complementary log-log regression                Number of obs     =     10,421
                                                Zero outcomes     =     10,371
                                                Nonzero outcomes  =         50

                                                Wald chi2(10)     =      74.20
Log pseudolikelihood = -263.52532               Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.522098    .164885     3.88   0.000     1.230932    1.882136
           gdppcthl |   1.804217   .2640801     4.03   0.000     1.354254    2.403686
          gdppcthl2 |   .9679587   .0095473    -3.30   0.001      .949426    .9868532
            polityl |   .9268404   .0307565    -2.29   0.022     .8684773    .9891256
           polityl2 |   .9840493   .0051954    -3.05   0.002      .973919     .994285
yrsincleaderinpower |   1.032795   .0125855     2.65   0.008      1.00842    1.057759
        v2x_execorr |   5.246124   4.189434     2.08   0.038     1.096702    25.09507
             lnoill |   .8773826   .0298906    -3.84   0.000     .8207111    .9379673
        postcoldwar |   6.753469   2.875851     4.49   0.000     2.931278    15.55954
         percurbanl |   1.001049   .0087029     0.12   0.904     .9841362    1.018253
              _cons |   5.60e-06   8.22e-06    -8.24   0.000     3.15e-07    .0000994
-------------------------------------------------------------------------------------

. *  Result  
. *    --percurbanl is statistically insignificant
. *    --no sign change or change in statistical significance for all other variables
. * Urbanization is significant when gdppcthl is dropped due to high correlation between the two (r=.68)
. cloglog urbancivicny lnpopl  polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldwar percurbanl if i
> ndstate==1, vce(robust) eform nolog

Complementary log-log regression                Number of obs     =     10,887
                                                Zero outcomes     =     10,837
                                                Nonzero outcomes  =         50

                                                Wald chi2(8)      =      62.59
Log pseudolikelihood = -271.18246               Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.601348    .180299     4.18   0.000     1.284243    1.996752
            polityl |   .9232448   .0315287    -2.34   0.019     .8634724     .987155
           polityl2 |   .9846334   .0049246    -3.10   0.002     .9750286    .9943329
yrsincleaderinpower |   1.027622    .012095     2.31   0.021     1.004187    1.051603
        v2x_execorr |   4.253801   3.254018     1.89   0.058     .9498083    19.05103
             lnoill |   .9080004   .0299803    -2.92   0.003      .851101    .9687037
        postcoldwar |   6.363599   2.734185     4.31   0.000     2.741419    14.77169
         percurbanl |    1.02456   .0096048     2.59   0.010     1.005906    1.043558
              _cons |   4.87e-06   7.54e-06    -7.89   0.000     2.33e-07    .0001016
-------------------------------------------------------------------------------------

. 
. * Population density: lnpopdensityl
. * Bivariate, controlling for time
. xtcloglog urbancivicny lnpopdensityl time1 timesq timecub if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,742
Group variable: cowcode                         Number of groups  =        164

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
                                                              avg =       71.6
                                                              max =        115

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      33.36
Log pseudolikelihood  = -305.08382              Prob > chi2       =     0.0000

                               (Std. Err. adjusted for 164 clusters in cowcode)
-------------------------------------------------------------------------------
              |               Robust
 urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------+----------------------------------------------------------------
lnpopdensityl |   1.257399   .1290367     2.23   0.026     1.028303    1.537535
        time1 |   .9468991    .075459    -0.68   0.494     .8099736    1.106972
       timesq |   1.001579   .0013267     1.19   0.234     .9989821    1.004183
      timecub |   .9999922   6.65e-06    -1.17   0.244     .9999792    1.000005
        _cons |   .0002425   .0004099    -4.92   0.000     8.83e-06     .006661
--------------+----------------------------------------------------------------
     /lnsig2u |  -1.954086   1.689017                     -5.264498    1.356326
--------------+----------------------------------------------------------------
      sigma_u |   .3764225    .317892                      .0719165    1.970255
          rho |    .079308    .123329                      .0031343    .7023734
-------------------------------------------------------------------------------

. * Result:  positive and statistically significant in bivariate relationship, controlling for time
. * Multivariate cloglog panel, Model 4 from Table 3.1
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar lnpopdensityl if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,516
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       67.0
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      97.01
Log pseudolikelihood  = -263.04523              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 157 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.446015   .1414679     3.77   0.000     1.193705    1.751654
           gdppcthl |   1.767041   .2123614     4.74   0.000     1.396208    2.236368
          gdppcthl2 |   .9691344    .007969    -3.81   0.000     .9536406      .98488
            polityl |   .9272299   .0306583    -2.29   0.022     .8690463     .989309
           polityl2 |   .9837056   .0055993    -2.89   0.004     .9727921    .9947416
yrsincleaderinpower |   1.033548   .0127456     2.68   0.007     1.008867    1.058833
        v2x_execorr |   5.149315    4.42188     1.91   0.056      .956744    27.71426
             lnoill |    .892814    .029765    -3.40   0.001     .8363408    .9531004
        postcoldwar |   6.430312   2.887754     4.14   0.000     2.666678    15.50578
      lnpopdensityl |   1.168286   .1705093     1.07   0.287     .8776424    1.555179
              _cons |   5.34e-06   7.25e-06    -8.94   0.000     3.73e-07    .0000765
--------------------+----------------------------------------------------------------
           /lnsig2u |   -11.0056   25953.72                     -50879.37    50857.36
--------------------+----------------------------------------------------------------
            sigma_u |   .0040753   52.88524                             0           .
                rho |   .0000101   .2620427                             0           .
-------------------------------------------------------------------------------------

. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -263.04523     -263.04521     -263.04521
likelihood                   .00001916      .00001916   Difference
                            -7.286e-08     -7.286e-08   Relative difference
-----------------------------------------------------
urbancivicny: .36881142      .36881404      .36881404
  lnpopl                     2.621e-06      2.621e-06   Difference
                             7.105e-06      7.105e-06   Relative difference
-----------------------------------------------------
urbancivicny: .56930658      .56929921      .56929921
gdppcthl                    -7.370e-06     -7.370e-06   Difference
                            -.00001294     -.00001294   Relative difference
-----------------------------------------------------
urbancivicny:-.03135197     -.03135066     -.03135066
gdppcthl2                    1.314e-06      1.314e-06   Difference
                            -.00004191     -.00004191   Relative difference
-----------------------------------------------------
urbancivicny: -.0755537     -.07554516     -.07554516
 polityl                     8.541e-06      8.541e-06   Difference
                            -.00011304     -.00011304   Relative difference
-----------------------------------------------------
urbancivicny:-.01642857     -.01642893     -.01642893
polityl2                    -3.595e-07     -3.595e-07   Difference
                             .00002188      .00002188   Relative difference
-----------------------------------------------------
urbancivicny: .03299766       .0329898       .0329898
yrsinclead~r                -7.860e-06     -7.860e-06   Difference
                             -.0002382      -.0002382   Relative difference
-----------------------------------------------------
urbancivicny: 1.6388637      1.6390703      1.6390703
v2x_execorr                  .00020661      .00020661   Difference
                             .00012607      .00012607   Relative difference
-----------------------------------------------------
urbancivicny:-.11337704     -.11337319     -.11337319
  lnoill                     3.852e-06      3.852e-06   Difference
                            -.00003397     -.00003397   Relative difference
-----------------------------------------------------
urbancivicny: 1.8610231      1.8609246      1.8609246
postcoldwar                 -.00009853     -.00009853   Difference
                            -.00005295     -.00005295   Relative difference
-----------------------------------------------------
urbancivicny: .15553737      .15554167      .15554167
lnpopdensi~l                 4.294e-06      4.294e-06   Difference
                             .00002761      .00002761   Relative difference
-----------------------------------------------------
urbancivicny:-12.140224      -12.13997      -12.13997
   _cons                     .00025462      .00025462   Difference
                            -.00002097     -.00002097   Relative difference
-----------------------------------------------------
lnsig2u:     -11.005598     -13.885605     -13.885605
   _cons                    -2.8800076     -2.8800077   Difference
                             .26168571      .26168571   Relative difference
-----------------------------------------------------

. * Failed quadchk--recalculate using pooled model
. cloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldw
> ar lnpopdensityl if indstate==1, vce(robust) eform nolog

Complementary log-log regression                Number of obs     =     10,516
                                                Zero outcomes     =     10,466
                                                Nonzero outcomes  =         50

                                                Wald chi2(10)     =      76.17
Log pseudolikelihood = -263.04521               Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.446016   .1716687     3.11   0.002     1.145829    1.824847
           gdppcthl |   1.767036   .2356873     4.27   0.000     1.360543    2.294979
          gdppcthl2 |   .9691349   .0085746    -3.54   0.000     .9524739    .9860873
            polityl |   .9272327   .0307582    -2.28   0.023     .8688657    .9895206
           polityl2 |   .9837055   .0052464    -3.08   0.002     .9734764    .9940421
yrsincleaderinpower |   1.033545   .0121086     2.82   0.005     1.010083    1.057552
        v2x_execorr |   5.149685   4.020113     2.10   0.036     1.115034    23.78335
             lnoill |   .8928152   .0351818    -2.88   0.004     .8264556     .964503
        postcoldwar |   6.430092   2.741057     4.37   0.000     2.788444    14.82765
      lnpopdensityl |   1.168287   .1894085     0.96   0.337     .8502541    1.605279
              _cons |   5.34e-06   7.60e-06    -8.53   0.000     3.28e-07    .0000869
-------------------------------------------------------------------------------------

. *  Result 
. *    --lnpopdensityl is statistically insignificant 
. *    --no sign change or change in statistical significance for all other variables
. 
. * Youth as proportion of population: youthpercl
. * Bivariate, controlling for time
. xtcloglog urbancivicny youthpercl time1 timesq timecub if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =      8,527
Group variable: cowcode                         Number of groups  =        159

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         22
                                                              avg =       53.6
                                                              max =         64

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      16.94
Log pseudolikelihood  = -286.35358              Prob > chi2       =     0.0020

                              (Std. Err. adjusted for 159 clusters in cowcode)
------------------------------------------------------------------------------
             |               Robust
urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
  youthpercl |   1.017022   .0185508     0.93   0.355     .9813053    1.054039
       time1 |   .9552909    .588357    -0.07   0.941     .2856844    3.194367
      timesq |   1.001204   .0070343     0.17   0.864     .9875118    1.015087
     timecub |   .9999948   .0000265    -0.20   0.843     .9999429    1.000047
       _cons |   .0004627   .0082267    -0.43   0.666     3.40e-19    6.29e+11
-------------+----------------------------------------------------------------
    /lnsig2u |  -1.342937   .9978627                     -3.298712    .6128377
-------------+----------------------------------------------------------------
     sigma_u |   .5109577   .2549328                      .1921736    1.358551
         rho |   .1369759   .1179609                      .0219582    .5287526
------------------------------------------------------------------------------

. * Result:  positive and statistically insignificant in bivariate relationship, controlling for time
. * Multivariate cloglog panel, Model 4 from Table 3.1
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar youthpercl if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =      8,111
Group variable: cowcode                         Number of groups  =        154

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       52.7
                                                              max =         64

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      91.54
Log pseudolikelihood  =  -242.6724              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 154 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.520115   .1566097     4.06   0.000     1.242172    1.860251
           gdppcthl |   1.589385   .2468546     2.98   0.003     1.172263     2.15493
          gdppcthl2 |   .9715734   .0093737    -2.99   0.003      .953374    .9901203
            polityl |   .9213099    .031802    -2.37   0.018     .8610409    .9857974
           polityl2 |   .9791664    .005994    -3.44   0.001     .9674887    .9909851
yrsincleaderinpower |   1.032727   .0142193     2.34   0.019      1.00523    1.060976
        v2x_execorr |   7.259798   6.788685     2.12   0.034     1.161343    45.38253
             lnoill |   .8881935   .0287825    -3.66   0.000     .8335351    .9464361
        postcoldwar |   4.677402   2.116475     3.41   0.001     1.926837    11.35441
         youthpercl |   .9474893   .0268652    -1.90   0.057      .896271    1.001635
              _cons |   .0000553   .0000998    -5.43   0.000     1.61e-06    .0018995
--------------------+----------------------------------------------------------------
           /lnsig2u |  -10.63202   20180.35                      -39563.4    39542.14
--------------------+----------------------------------------------------------------
            sigma_u |   .0049123   49.56604                             0           .
                rho |   .0000147   .2960318                             0           .
-------------------------------------------------------------------------------------

. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log           -242.6724     -242.67237     -242.67237
likelihood                   .00003175      .00003175   Difference
                            -1.309e-07     -1.309e-07   Relative difference
-----------------------------------------------------
urbancivicny: .41878627      .41877444      .41877444
  lnpopl                    -.00001183     -.00001183   Difference
                            -.00002824     -.00002824   Relative difference
-----------------------------------------------------
urbancivicny: .46334726      .46333245      .46333245
gdppcthl                    -.00001481     -.00001481   Difference
                            -.00003196     -.00003196   Relative difference
-----------------------------------------------------
urbancivicny:-.02883842     -.02883856     -.02883856
gdppcthl2                   -1.309e-07     -1.309e-07   Difference
                             4.539e-06      4.539e-06   Relative difference
-----------------------------------------------------
urbancivicny:-.08195886     -.08196905     -.08196905
 polityl                    -.00001018     -.00001018   Difference
                             .00012425      .00012425   Relative difference
-----------------------------------------------------
urbancivicny:-.02105366     -.02105376     -.02105376
polityl2                    -1.000e-07     -1.000e-07   Difference
                             4.752e-06      4.752e-06   Relative difference
-----------------------------------------------------
urbancivicny: .03220293       .0322128       .0322128
yrsinclead~r                 9.871e-06      9.871e-06   Difference
                             .00030654      .00030654   Relative difference
-----------------------------------------------------
urbancivicny:  1.982352      1.9821434      1.9821434
v2x_execorr                 -.00020867     -.00020867   Difference
                            -.00010526     -.00010526   Relative difference
-----------------------------------------------------
urbancivicny:-.11856563     -.11856439     -.11856439
  lnoill                     1.234e-06      1.234e-06   Difference
                            -.00001041     -.00001041   Relative difference
-----------------------------------------------------
urbancivicny: 1.5427428      1.5428394      1.5428394
postcoldwar                   .0000966       .0000966   Difference
                             .00006262      .00006262   Relative difference
-----------------------------------------------------
urbancivicny: -.0539396      -.0539435      -.0539435
youthpercl                  -3.904e-06     -3.904e-06   Difference
                             .00007237      .00007237   Relative difference
-----------------------------------------------------
urbancivicny:-9.8020837     -9.8021667     -9.8021667
   _cons                    -.00008302     -.00008302   Difference
                             8.470e-06      8.470e-06   Relative difference
-----------------------------------------------------
lnsig2u:     -10.632024      -13.49935      -13.49935
   _cons                    -2.8673264     -2.8673264   Difference
                             .26968774      .26968774   Relative difference
-----------------------------------------------------

. * Failed quadchk--recalculate using pooled model
. cloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldw
> ar youthpercl if indstate==1, vce(robust) eform nolog

Complementary log-log regression                Number of obs     =      8,111
                                                Zero outcomes     =      8,064
                                                Nonzero outcomes  =         47

                                                Wald chi2(10)     =      68.40
Log pseudolikelihood = -242.67237               Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.520109   .1816424     3.50   0.000     1.202713    1.921266
           gdppcthl |   1.589377   .2488757     2.96   0.003     1.169337    2.160299
          gdppcthl2 |   .9715734   .0097137    -2.88   0.004     .9527203    .9907996
            polityl |   .9213066   .0321254    -2.35   0.019     .8604454    .9864727
           polityl2 |   .9791664   .0057613    -3.58   0.000     .9679392    .9905238
yrsincleaderinpower |   1.032731   .0132854     2.50   0.012     1.007017    1.059101
        v2x_execorr |   7.259271    6.32651     2.27   0.023      1.31542    40.06099
             lnoill |   .8881939   .0313503    -3.36   0.001     .8288258    .9518146
        postcoldwar |   4.677559   2.109186     3.42   0.001     1.932854    11.31982
         youthpercl |    .947488   .0264917    -1.93   0.054     .8969623     1.00086
              _cons |   .0000553   .0001056    -5.14   0.000     1.31e-06    .0023302
-------------------------------------------------------------------------------------

. *  Result 
. *    --sample is reduced because data on youthpercl only begins in 1951 (n=8,111)
. *    --youthpercl is negative (opposite of what theory would say) and marginally significant at the .10 level
. *    --no sign change or change in statistical significance for all other variables
. 
. * Trade as percent of GDP: tottradepernomgdpl 
. * Bivariate, controlling for time
. xtcloglog urbancivicny tottradepernomgdpl time1 timesq timecub if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =      9,310
Group variable: cowcode                         Number of groups  =        155

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         13
                                                              avg =       60.1
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      16.64
Log pseudolikelihood  =  -267.8544              Prob > chi2       =     0.0023

                                    (Std. Err. adjusted for 155 clusters in cowcode)
------------------------------------------------------------------------------------
                   |               Robust
      urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------------+----------------------------------------------------------------
tottradepernomgdpl |   .9011295   .2078925    -0.45   0.652     .5733412    1.416319
             time1 |   .9392492   .0748268    -0.79   0.431     .8034678    1.097977
            timesq |   1.001482    .001421     1.04   0.297     .9987008    1.004271
           timecub |   .9999933   7.37e-06    -0.92   0.360     .9999788    1.000008
             _cons |    .001287   .0020078    -4.27   0.000     .0000605    .0273848
-------------------+----------------------------------------------------------------
          /lnsig2u |  -1.600983    1.40914                     -4.362847    1.160881
-------------------+----------------------------------------------------------------
           sigma_u |   .4491081   .3164282                      .1128807    1.786825
               rho |   .1092248    .137102                      .0076867    .6599745
------------------------------------------------------------------------------------

. * Result:  negative and statistically insignificant  in bivariate relationship, controlling for time
. * Multivariate cloglog panel, Model 4 from Table 3.1
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar tottradepernomgdpl if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =      8,941
Group variable: cowcode                         Number of groups  =        150

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         13
                                                              avg =       59.6
                                                              max =        113

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      67.13
Log pseudolikelihood  = -232.66896              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 150 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.515821    .160891     3.92   0.000      1.23112     1.86636
           gdppcthl |    1.84326   .3073524     3.67   0.000     1.329395    2.555756
          gdppcthl2 |   .9633932   .0111714    -3.22   0.001     .9417445    .9855395
            polityl |   .9259625   .0327386    -2.18   0.030     .8639689    .9924045
           polityl2 |   .9868385   .0056916    -2.30   0.022     .9757459    .9980572
yrsincleaderinpower |   1.024044   .0138642     1.75   0.079     .9972279    1.051581
        v2x_execorr |   6.097566   5.824717     1.89   0.058     .9376567    39.65237
             lnoill |   .8803553   .0318038    -3.53   0.000     .8201767    .9449493
        postcoldwar |    5.46203   2.306125     4.02   0.000     2.387636    12.49511
 tottradepernomgdpl |   .9815083   .0664018    -0.28   0.783     .8596224    1.120676
              _cons |   6.81e-06   .0000101    -8.03   0.000     3.74e-07    .0001241
--------------------+----------------------------------------------------------------
           /lnsig2u |  -10.13963   13290.36                     -26058.76    26038.48
--------------------+----------------------------------------------------------------
            sigma_u |   .0062836   41.75544                             0           .
                rho |    .000024   .3189923                             0           .
-------------------------------------------------------------------------------------

. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -232.66896     -232.66892     -232.66892
likelihood                   .00003976      .00004186   Difference
                            -1.709e-07     -1.799e-07   Relative difference
-----------------------------------------------------
urbancivicny: .41595741      .41592578      .41593166
  lnpopl                    -.00003163     -.00002575   Difference
                            -.00007604      -.0000619   Relative difference
-----------------------------------------------------
urbancivicny:   .611536      .61153254      .61153319
gdppcthl                    -3.459e-06     -2.816e-06   Difference
                            -5.657e-06     -4.605e-06   Relative difference
-----------------------------------------------------
urbancivicny:-.03729368     -.03729487     -.03729465
gdppcthl2                   -1.195e-06     -9.727e-07   Difference
                             .00003204      .00002608   Relative difference
-----------------------------------------------------
urbancivicny:-.07692149     -.07693404      -.0769317
 polityl                    -.00001255     -.00001021   Difference
                             .00016312      .00013279   Relative difference
-----------------------------------------------------
urbancivicny:-.01324888     -.01324971     -.01324956
polityl2                    -8.302e-07     -6.759e-07   Difference
                             .00006266      .00005101   Relative difference
-----------------------------------------------------
urbancivicny:  .0237594      .02376968      .02376777
yrsinclead~r                 .00001029      8.373e-06   Difference
                             .00043291      .00035241   Relative difference
-----------------------------------------------------
urbancivicny: 1.8078897      1.8074946      1.8075681
v2x_execorr                 -.00039514     -.00032166   Difference
                            -.00021856     -.00017792   Relative difference
-----------------------------------------------------
urbancivicny:-.12742975     -.12742394     -.12742502
  lnoill                     5.813e-06      4.732e-06   Difference
                            -.00004561     -.00003713   Relative difference
-----------------------------------------------------
urbancivicny: 1.6978206      1.6978878      1.6978753
postcoldwar                  .00006719       .0000547   Difference
                             .00003958      .00003222   Relative difference
-----------------------------------------------------
urbancivicny:-.01866486     -.01866255     -.01866298
tottradepe~l                 2.309e-06      1.880e-06   Difference
                            -.00012369      -.0001007   Relative difference
-----------------------------------------------------
urbancivicny:-11.897148      -11.89713     -11.897134
   _cons                     .00001794       .0000146   Difference
                            -1.508e-06     -1.227e-06   Relative difference
-----------------------------------------------------
lnsig2u:     -10.139634     -12.128681       -12.5055
   _cons                    -1.9890471     -2.3658658   Difference
                             .19616557      .23332852   Relative difference
-----------------------------------------------------

. *  Result 
. *    --sample is reduced because of missing data on tottradepernomgdpl (n=9,310), with only 43 urban civic episo
> des and 150 countries
. *    --tottradepernomgdpl is statistically insignificant 
. *    --yrsincleaderinpower and v2x_execorr grow marginally significant at .10 level
. *    --no sign change or change in statistical significance for all other variables
. 
. * Logged dollar exchange rate:  lndollexchratel 
. * Bivariate, controlling for time
. xtcloglog urbancivicny lndollexchratel time1 timesq timecub if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =      8,826
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          3
                                                              avg =       56.2
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      21.30
Log pseudolikelihood  = -263.21349              Prob > chi2       =     0.0003

                                 (Std. Err. adjusted for 157 clusters in cowcode)
---------------------------------------------------------------------------------
                |               Robust
   urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
lndollexchratel |   1.070681    .065281     1.12   0.263     .9500824    1.206589
          time1 |   .9324823     .07856    -0.83   0.407     .7905483    1.099899
         timesq |   1.001305   .0013867     0.94   0.347     .9985903    1.004026
        timecub |    .999995   6.88e-06    -0.73   0.466     .9999815    1.000008
          _cons |   .0017802   .0031157    -3.62   0.000     .0000576    .0549845
----------------+----------------------------------------------------------------
       /lnsig2u |  -.9347804   .7106071                     -2.327545     .457984
----------------+----------------------------------------------------------------
        sigma_u |   .6266355   .2226458                      .3123058    1.257332
            rho |   .1927125   .1105523                      .0559751    .4900722
---------------------------------------------------------------------------------

. * Result:  positive and statistically insignificant in bivariate relationship, controlling for time
. * Multivariate cloglog panel, Model 4 from Table 3.1
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar lndollexchratel if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =      8,455
Group variable: cowcode                         Number of groups  =        150

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =          6
                                                              avg =       56.4
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      83.53
Log pseudolikelihood  = -221.80978              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 150 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.603057   .1544958     4.90   0.000     1.327131    1.936351
           gdppcthl |   2.054207   .3531792     4.19   0.000     1.466551    2.877339
          gdppcthl2 |   .9568569   .0126349    -3.34   0.001     .9324106    .9819442
            polityl |   .9391928   .0327854    -1.80   0.072     .8770835      1.0057
           polityl2 |   .9906228   .0060861    -1.53   0.125     .9787657    1.002623
yrsincleaderinpower |   1.043605   .0146423     3.04   0.002     1.015297    1.072701
        v2x_execorr |   20.67751   17.90218     3.50   0.000      3.78918     112.837
             lnoill |   .8623863   .0300927    -4.24   0.000     .8053774    .9234307
        postcoldwar |   5.236605   2.486387     3.49   0.000     2.064866     13.2803
    lndollexchratel |   1.010282   .0688474     0.15   0.881     .8839674    1.154647
              _cons |   1.04e-06   1.50e-06    -9.55   0.000     6.12e-08    .0000175
--------------------+----------------------------------------------------------------
           /lnsig2u |  -4.105003   23.44746                     -50.06118    41.85117
--------------------+----------------------------------------------------------------
            sigma_u |   .1284133   1.505482                      1.35e-11    1.22e+09
                rho |   .0099252    .230411                      1.10e-22           1
-------------------------------------------------------------------------------------

. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -221.80978     -221.80978     -221.80978
likelihood                   1.923e-07      1.923e-07   Difference
                            -8.667e-10     -8.667e-10   Relative difference
-----------------------------------------------------
urbancivicny: .47191246      .47192001      .47192001
  lnpopl                     7.551e-06      7.551e-06   Difference
                               .000016        .000016   Relative difference
-----------------------------------------------------
urbancivicny: .71988973      .71988799      .71988799
gdppcthl                    -1.744e-06     -1.744e-06   Difference
                            -2.422e-06     -2.422e-06   Relative difference
-----------------------------------------------------
urbancivicny: -.0441014     -.04410091     -.04410091
gdppcthl2                    4.900e-07      4.900e-07   Difference
                            -.00001111     -.00001111   Relative difference
-----------------------------------------------------
urbancivicny:-.06273453     -.06273061     -.06273061
 polityl                     3.920e-06      3.920e-06   Difference
                            -.00006249     -.00006249   Relative difference
-----------------------------------------------------
urbancivicny:-.00942148     -.00942128     -.00942128
polityl2                     2.041e-07      2.041e-07   Difference
                            -.00002166     -.00002166   Relative difference
-----------------------------------------------------
urbancivicny: .04268073      .04267686      .04267686
yrsinclead~r                -3.876e-06     -3.876e-06   Difference
                            -.00009082     -.00009082   Relative difference
-----------------------------------------------------
urbancivicny: 3.0290469      3.0290884      3.0290884
v2x_execorr                  .00004155      .00004155   Difference
                             .00001372      .00001372   Relative difference
-----------------------------------------------------
urbancivicny:-.14805192     -.14805577     -.14805577
  lnoill                    -3.855e-06     -3.855e-06   Difference
                             .00002604      .00002604   Relative difference
-----------------------------------------------------
urbancivicny: 1.6556734      1.6556558      1.6556558
postcoldwar                 -.00001762     -.00001762   Difference
                            -.00001064     -.00001064   Relative difference
-----------------------------------------------------
urbancivicny: .01022994      .01023129      .01023129
lndollexch~l                 1.353e-06      1.353e-06   Difference
                              .0001323       .0001323   Relative difference
-----------------------------------------------------
urbancivicny:-13.780144     -13.780049     -13.780049
   _cons                     .00009504      .00009504   Difference
                            -6.897e-06     -6.897e-06   Relative difference
-----------------------------------------------------
lnsig2u:     -4.1050031     -4.1205478     -4.1205478
   _cons                    -.01554474     -.01554473   Difference
                             .00378678      .00378678   Relative difference
-----------------------------------------------------

. *  Result 
. *    --sample is highly reducted (n=8,455)
. *    --lndollexchratel is statistically insignificant
. *    --polity grows marginally significant at the .10 level, and polityl2 becomes statistically insignificant
. *    --no sign change or change in statistical significance for all other variables
. 
. * Presence of financial crisis: rrfinstressl
. * Bivariate, controlling for time
. xtcloglog urbancivicny rrfinstressl time1 timesq timecub if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =      6,319
Group variable: cowcode                         Number of groups  =         68

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         37
                                                              avg =       92.9
                                                              max =        111

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =       8.65
Log pseudolikelihood  = -122.45495              Prob > chi2       =     0.0703

                               (Std. Err. adjusted for 68 clusters in cowcode)
------------------------------------------------------------------------------
             |               Robust
urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
rrfinstressl |    1.03612    .246415     0.15   0.881     .6500892    1.651379
       time1 |   .9446754   .0798596    -0.67   0.501      .800433    1.114911
      timesq |   1.001771   .0016831     1.05   0.292     .9984772    1.005075
     timecub |   .9999902   9.54e-06    -1.02   0.306     .9999715    1.000009
       _cons |   .0005531   .0008544    -4.85   0.000     .0000268     .011423
-------------+----------------------------------------------------------------
    /lnsig2u |  -.7222647   .9319663                     -2.548885    1.104356
-------------+----------------------------------------------------------------
     sigma_u |   .6968868   .3247375                      .2795868    1.737032
         rho |   .2279426   .1640119                      .0453651    .6471775
------------------------------------------------------------------------------

. * Result:  positive and statistically insignificant in bivariate relationship, controlling for time
. * Multivariate cloglog panel, Model 4 from Table 3.1
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar rrfinstressl if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =      5,887
Group variable: cowcode                         Number of groups  =         67

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         36
                                                              avg =       87.9
                                                              max =        111

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      88.07
Log pseudolikelihood  = -99.700843              Prob > chi2       =     0.0000

                                      (Std. Err. adjusted for 67 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.477598   .2316562     2.49   0.013     1.086691    2.009124
           gdppcthl |   2.799558   .8030736     3.59   0.000     1.595571    4.912051
          gdppcthl2 |     .93281   .0230474    -2.82   0.005     .8887143    .9790936
            polityl |   .8812119    .051461    -2.17   0.030     .7859084    .9880726
           polityl2 |   .9917685   .0098616    -0.83   0.406     .9726273    1.011286
yrsincleaderinpower |   1.022568   .0194561     1.17   0.241     .9851368    1.061421
        v2x_execorr |    32.2791   36.31766     3.09   0.002     3.558139    292.8329
             lnoill |   .9023739   .0440181    -2.11   0.035     .8200958    .9929067
        postcoldwar |   4.032789   2.077574     2.71   0.007     1.469236    11.06928
       rrfinstressl |   .9623199   .2081809    -0.18   0.859     .6297628    1.470489
              _cons |   7.70e-07   1.56e-06    -6.94   0.000     1.44e-08    .0000411
--------------------+----------------------------------------------------------------
           /lnsig2u |  -12.97288          .                             .           .
--------------------+----------------------------------------------------------------
            sigma_u |    .001524          .                             .           .
                rho |   1.41e-06          .                             .           .
-------------------------------------------------------------------------------------

. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -99.700843     -99.700834     -99.700834
likelihood                   8.843e-06      8.843e-06   Difference
                            -8.869e-08     -8.869e-08   Relative difference
-----------------------------------------------------
urbancivicny: .39041781      .39029879      .39029879
  lnpopl                    -.00011902     -.00011902   Difference
                            -.00030486     -.00030486   Relative difference
-----------------------------------------------------
urbancivicny: 1.0294615      1.0294719      1.0294719
gdppcthl                     .00001032      .00001032   Difference
                             .00001003      .00001003   Relative difference
-----------------------------------------------------
urbancivicny:-.06955372     -.06955865     -.06955865
gdppcthl2                   -4.931e-06     -4.931e-06   Difference
                             .00007089      .00007089   Relative difference
-----------------------------------------------------
urbancivicny:-.12645711     -.12644603     -.12644603
 polityl                     .00001108      .00001108   Difference
                            -.00008763     -.00008763   Relative difference
-----------------------------------------------------
urbancivicny:-.00826552     -.00826623     -.00826623
polityl2                    -7.032e-07     -7.032e-07   Difference
                             .00008508      .00008508   Relative difference
-----------------------------------------------------
urbancivicny: .02231694       .0223251       .0223251
yrsinclead~r                 8.157e-06      8.157e-06   Difference
                             .00036553      .00036553   Relative difference
-----------------------------------------------------
urbancivicny: 3.4744199      3.4739056      3.4739056
v2x_execorr                 -.00051436     -.00051436   Difference
                            -.00014804     -.00014804   Relative difference
-----------------------------------------------------
urbancivicny:-.10272633     -.10268847     -.10268847
  lnoill                     .00003786      .00003786   Difference
                            -.00036858     -.00036858   Relative difference
-----------------------------------------------------
urbancivicny: 1.3944582      1.3943174      1.3943174
postcoldwar                 -.00014085     -.00014085   Difference
                            -.00010101     -.00010101   Relative difference
-----------------------------------------------------
urbancivicny:-.03840834     -.03837135     -.03837135
rrfinstressl                 .00003699      .00003699   Difference
                            -.00096312     -.00096312   Relative difference
-----------------------------------------------------
urbancivicny:-14.077371     -14.076293     -14.076293
   _cons                     .00107762      .00107762   Difference
                            -.00007655     -.00007655   Relative difference
-----------------------------------------------------
lnsig2u:     -12.972882     -14.345926     -14.345926
   _cons                    -1.3730436     -1.3730436   Difference
                             .10583952      .10583952   Relative difference
-----------------------------------------------------

. * Does not pass quadchk--recalculate using pooled model
. cloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldw
> ar rrfinstressl if indstate==1, vce(robust) eform nolog

Complementary log-log regression                Number of obs     =      5,887
                                                Zero outcomes     =      5,868
                                                Nonzero outcomes  =         19

                                                Wald chi2(10)     =      35.57
Log pseudolikelihood = -99.700832               Prob > chi2       =     0.0001

-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |    1.47747   .2842763     2.03   0.042     1.013307    2.154251
           gdppcthl |   2.799579   .9850481     2.93   0.003     1.404729     5.57947
          gdppcthl2 |   .9328067   .0285384    -2.27   0.023     .8785164    .9904519
            polityl |    .881219   .0550219    -2.03   0.043     .7797156    .9959362
           polityl2 |    .991768   .0099423    -0.82   0.410     .9724717    1.011447
yrsincleaderinpower |   1.022574   .0232116     0.98   0.325     .9780771    1.069095
        v2x_execorr |   32.26702   41.04327     2.73   0.006     2.667109    390.3706
             lnoill |   .9023987    .051556    -1.80   0.072     .8068029    1.009321
        postcoldwar |   4.032376   2.224648     2.53   0.011     1.367591    11.88956
       rrfinstressl |   .9623458   .1936148    -0.19   0.849     .6487502    1.427529
              _cons |   7.70e-07   1.80e-06    -6.04   0.000     7.97e-09    .0000745
-------------------------------------------------------------------------------------

. *  Result 
. *    --sample is highly reduced because of missing data on tradepergdpl (n=5,887) with 19 urban civic episodes a
> nd 67 countries
. *    --rrfinstressl is statistically insignificant
. *    --polityl2 and yrsincleaderinpower grow statistically insignificant
. *    --no sign change or change in statistical significance for all other variables
. 
. *  Child mortality:  under5mortl
. * Bivariate, controlling for time
. xtcloglog urbancivicny under5mortl time1 timesq timecub if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,537
Group variable: cowcode                         Number of groups  =        159

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         22
                                                              avg =       72.6
                                                              max =        115

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      30.83
Log pseudolikelihood  = -295.83097              Prob > chi2       =     0.0000

                              (Std. Err. adjusted for 159 clusters in cowcode)
------------------------------------------------------------------------------
             |               Robust
urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
 under5mortl |   .9982437   .0019249    -0.91   0.362     .9944781    1.002024
       time1 |    .958165   .0789565    -0.52   0.604     .8152634    1.126115
      timesq |   1.001251   .0013473     0.93   0.353     .9986134    1.003895
     timecub |   .9999942   6.69e-06    -0.87   0.383      .999981    1.000007
       _cons |    .000747   .0014898    -3.61   0.000      .000015    .0372356
-------------+----------------------------------------------------------------
    /lnsig2u |  -1.456446   1.082795                     -3.578685    .6657935
-------------+----------------------------------------------------------------
     sigma_u |   .4827662   .2613684                        .16707    1.395003
         rho |    .124102   .1177006                      .0166856    .5419246
------------------------------------------------------------------------------

. * Result:  negative and statistically insignificant in bivariate relationship, controlling for time
. * Multivariate cloglog panel, Model 4 from Table 3.1
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar under5mortl if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,355
Group variable: cowcode                         Number of groups  =        154

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       67.2
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      82.86
Log pseudolikelihood  =  -255.0694              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 154 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.484726   .1483977     3.95   0.000     1.220588    1.806024
           gdppcthl |   1.587072   .2438918     3.01   0.003     1.174327    2.144886
          gdppcthl2 |   .9726204   .0091475    -2.95   0.003     .9548559    .9907154
            polityl |    .929496    .031685    -2.14   0.032     .8694237     .993719
           polityl2 |   .9832012   .0057981    -2.87   0.004     .9719027    .9946312
yrsincleaderinpower |   1.028801    .013646     2.14   0.032       1.0024    1.055898
        v2x_execorr |   6.057373   5.337391     2.04   0.041     1.077102    34.06525
             lnoill |   .8900747     .02843    -3.65   0.000     .8360613    .9475776
        postcoldwar |   4.293506   2.051575     3.05   0.002     1.682987    10.95326
        under5mortl |   .9955264   .0025216    -1.77   0.077     .9905964    1.000481
              _cons |   .0000208   .0000321    -6.99   0.000     1.01e-06    .0004283
--------------------+----------------------------------------------------------------
           /lnsig2u |  -10.62306   15085.82                     -29578.28    29557.04
--------------------+----------------------------------------------------------------
            sigma_u |   .0049344   37.21955                             0           .
                rho |   .0000148   .2232914                             0           .
-------------------------------------------------------------------------------------

. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log           -255.0694     -255.06938     -255.06938
likelihood                   .00001728      .00001728   Difference
                            -6.777e-08     -6.777e-08   Relative difference
-----------------------------------------------------
urbancivicny: .39523043      .39521088      .39521088
  lnpopl                    -.00001955     -.00001955   Difference
                            -.00004946     -.00004946   Relative difference
-----------------------------------------------------
urbancivicny: .46189087      .46187924      .46187924
gdppcthl                    -.00001163     -.00001163   Difference
                            -.00002517     -.00002517   Relative difference
-----------------------------------------------------
urbancivicny:-.02776142     -.02776183     -.02776183
gdppcthl2                   -4.105e-07     -4.105e-07   Difference
                             .00001479      .00001479   Relative difference
-----------------------------------------------------
urbancivicny:-.07311278     -.07312193     -.07312193
 polityl                    -9.143e-06     -9.143e-06   Difference
                             .00012505      .00012505   Relative difference
-----------------------------------------------------
urbancivicny:-.01694146     -.01694135     -.01694135
polityl2                     1.141e-07      1.141e-07   Difference
                            -6.733e-06     -6.733e-06   Relative difference
-----------------------------------------------------
urbancivicny:  .0283941       .0284027       .0284027
yrsinclead~r                 8.598e-06      8.598e-06   Difference
                              .0003028       .0003028   Relative difference
-----------------------------------------------------
urbancivicny: 1.8012762      1.8010905      1.8010905
v2x_execorr                 -.00018564     -.00018564   Difference
                            -.00010306     -.00010306   Relative difference
-----------------------------------------------------
urbancivicny:-.11644985     -.11645156     -.11645156
  lnoill                    -1.709e-06     -1.709e-06   Difference
                             .00001467      .00001467   Relative difference
-----------------------------------------------------
urbancivicny: 1.4571037      1.4571541      1.4571541
postcoldwar                  .00005037      .00005037   Difference
                             .00003457      .00003457   Relative difference
-----------------------------------------------------
urbancivicny:-.00448368     -.00448426     -.00448426
under5mortl                 -5.856e-07     -5.856e-07   Difference
                              .0001306       .0001306   Relative difference
-----------------------------------------------------
urbancivicny: -10.77951       -10.7795       -10.7795
   _cons                     9.820e-06      9.820e-06   Difference
                            -9.110e-07     -9.110e-07   Relative difference
-----------------------------------------------------
lnsig2u:     -10.623058     -12.482866     -12.482866
   _cons                    -1.8598078     -1.8598078   Difference
                             .17507274      .17507274   Relative difference
-----------------------------------------------------

. * Does not pass quadchk--recalculate using pooled model
. cloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldw
> ar under5mortl if indstate==1, vce(robust) eform nolog

Complementary log-log regression                Number of obs     =     10,355
                                                Zero outcomes     =     10,307
                                                Nonzero outcomes  =         48

                                                Wald chi2(10)     =      73.32
Log pseudolikelihood = -255.06938               Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.484711   .1705262     3.44   0.001     1.185434    1.859544
           gdppcthl |   1.587062   .2599104     2.82   0.005     1.151315     2.18773
          gdppcthl2 |   .9726202   .0098242    -2.75   0.006     .9535544    .9920672
            polityl |   .9294915   .0320019    -2.12   0.034     .8688384    .9943787
           polityl2 |   .9832013     .00548    -3.04   0.002     .9725192    .9940007
yrsincleaderinpower |   1.028806   .0133333     2.19   0.028     1.003002    1.055273
        v2x_execorr |   6.056773   4.913762     2.22   0.026     1.235023    29.70348
             lnoill |   .8900739    .031775    -3.26   0.001     .8299249    .9545823
        postcoldwar |   4.293621   2.153868     2.90   0.004      1.60629    11.47687
        under5mortl |    .995526   .0025223    -1.77   0.077     .9905946    1.000482
              _cons |   .0000208    .000034    -6.61   0.000     8.51e-07    .0005094
-------------------------------------------------------------------------------------

. *  Result 
. *    --under5mortl is negative and marginally significant at the .10 level
. *    --no sign change or change in statistical significance for all other variables
. 
. * Age of incumbent leader: ageincleader
. * Bivariate, controlling for time
. xtcloglog urbancivicny ageincleader time1 timesq timecub if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,651
Group variable: cowcode                         Number of groups  =        164

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
                                                              avg =       71.0
                                                              max =        115

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      37.12
Log pseudolikelihood  = -302.27288              Prob > chi2       =     0.0000

                              (Std. Err. adjusted for 164 clusters in cowcode)
------------------------------------------------------------------------------
             |               Robust
urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
ageincleader |   1.039878   .0135262     3.01   0.003     1.013703     1.06673
       time1 |    .948284   .0742057    -0.68   0.497     .8134477    1.105471
      timesq |   1.001578   .0013024     1.21   0.225     .9990283    1.004134
     timecub |   .9999922   6.53e-06    -1.19   0.233     .9999794    1.000005
       _cons |   .0000506   .0000944    -5.30   0.000     1.31e-06    .0019594
-------------+----------------------------------------------------------------
    /lnsig2u |  -1.331257   .9242479                      -3.14275    .4802351
-------------+----------------------------------------------------------------
     sigma_u |   .5139503   .2375087                      .2077593    1.271399
         rho |   .1383625   .1101873                      .0255696    .4956338
------------------------------------------------------------------------------

. * Result:  positive and statistically significant in bivariate relationship, controlling for time
. * Multivariate cloglog panel, Model 4 from Table 3.1
. xtcloglog urbancivicny ageincleader time1 timesq timecub if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     11,651
Group variable: cowcode                         Number of groups  =        164

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
                                                              avg =       71.0
                                                              max =        115

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(4)      =      37.12
Log pseudolikelihood  = -302.27288              Prob > chi2       =     0.0000

                              (Std. Err. adjusted for 164 clusters in cowcode)
------------------------------------------------------------------------------
             |               Robust
urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
ageincleader |   1.039878   .0135262     3.01   0.003     1.013703     1.06673
       time1 |    .948284   .0742057    -0.68   0.497     .8134477    1.105471
      timesq |   1.001578   .0013024     1.21   0.225     .9990283    1.004134
     timecub |   .9999922   6.53e-06    -1.19   0.233     .9999794    1.000005
       _cons |   .0000506   .0000944    -5.30   0.000     1.31e-06    .0019594
-------------+----------------------------------------------------------------
    /lnsig2u |  -1.331257   .9242479                      -3.14275    .4802351
-------------+----------------------------------------------------------------
     sigma_u |   .5139503   .2375087                      .2077593    1.271399
         rho |   .1383625   .1101873                      .0255696    .4956338
------------------------------------------------------------------------------

. * Result:  positive and statistically signficant in bivariate relationship, controlling for time
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar ageincleader if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,506
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       66.9
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =     105.94
Log pseudolikelihood  = -263.01675              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 157 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.473954   .1499327     3.81   0.000     1.207531    1.799158
           gdppcthl |   1.812074    .238164     4.52   0.000     1.400558    2.344503
          gdppcthl2 |   .9676281   .0087637    -3.63   0.000     .9506032    .9849579
            polityl |   .9247697   .0307695    -2.35   0.019     .8663868    .9870867
           polityl2 |   .9837359   .0055412    -2.91   0.004     .9729352    .9946565
yrsincleaderinpower |   1.024639   .0134511     1.85   0.064     .9986117    1.051345
        v2x_execorr |   5.044055   4.324269     1.89   0.059     .9398199    27.07167
             lnoill |   .8776232   .0259852    -4.41   0.000     .8281428      .93006
        postcoldwar |   6.810804   2.981213     4.38   0.000     2.888098    16.06145
       ageincleader |    1.01577   .0133112     1.19   0.232     .9900126    1.042197
              _cons |   3.43e-06   4.47e-06    -9.67   0.000     2.68e-07     .000044
--------------------+----------------------------------------------------------------
           /lnsig2u |  -10.23365   12220.48                     -23961.94    23941.48
--------------------+----------------------------------------------------------------
            sigma_u |    .005995   36.63102                             0           .
                rho |   .0000218   .2669944                             0           .
-------------------------------------------------------------------------------------

. quadchk, nooutput

Refitting model intpoints() =  8
Refitting model intpoints() = 16

                         Quadrature check

               Fitted       Comparison     Comparison
             quadrature     quadrature     quadrature
             12 points      8 points       16 points
-----------------------------------------------------
Log          -263.01675     -263.01672     -263.01672
likelihood                   .00003151      .00003151   Difference
                            -1.198e-07     -1.198e-07   Relative difference
-----------------------------------------------------
urbancivicny:  .3879484      .38792942      .38792942
  lnpopl                    -.00001898     -.00001898   Difference
                            -.00004894     -.00004894   Relative difference
-----------------------------------------------------
urbancivicny: .59447205      .59446784      .59446784
gdppcthl                    -4.203e-06     -4.203e-06   Difference
                            -7.069e-06     -7.069e-06   Relative difference
-----------------------------------------------------
urbancivicny:-.03290746      -.0329085      -.0329085
gdppcthl2                   -1.039e-06     -1.039e-06   Difference
                             .00003158      .00003158   Relative difference
-----------------------------------------------------
urbancivicny: -.0782106     -.07822606     -.07822606
 polityl                    -.00001546     -.00001546   Difference
                             .00019769      .00019769   Relative difference
-----------------------------------------------------
urbancivicny: -.0163978     -.01639769     -.01639769
polityl2                     1.168e-07      1.168e-07   Difference
                            -7.123e-06     -7.123e-06   Relative difference
-----------------------------------------------------
urbancivicny: .02434048      .02435212      .02435212
yrsinclead~r                 .00001165      .00001165   Difference
                             .00047844      .00047844   Relative difference
-----------------------------------------------------
urbancivicny: 1.6182103      1.6178826      1.6178826
v2x_execorr                 -.00032773     -.00032773   Difference
                            -.00020253     -.00020253   Relative difference
-----------------------------------------------------
urbancivicny:-.13053792      -.1305389      -.1305389
  lnoill                    -9.781e-07     -9.781e-07   Difference
                             7.493e-06      7.493e-06   Relative difference
-----------------------------------------------------
urbancivicny: 1.9185102      1.9186566      1.9186566
postcoldwar                  .00014636      .00014636   Difference
                             .00007629      .00007629   Relative difference
-----------------------------------------------------
urbancivicny: .01564679      .01565152      .01565152
ageincleader                 4.730e-06      4.730e-06   Difference
                             .00030233      .00030233   Relative difference
-----------------------------------------------------
urbancivicny:-12.581928     -12.582518     -12.582518
   _cons                    -.00058998     -.00058998   Difference
                             .00004689      .00004689   Relative difference
-----------------------------------------------------
lnsig2u:     -10.233653     -12.092652     -12.092652
   _cons                    -1.8589992     -1.8589992   Difference
                             .18165549      .18165549   Relative difference
-----------------------------------------------------

. *  Result 
. *    --ageincleader is statistically insignificant 
. *    --yrsincleaderinpower grows statistically insignificant 
. *    --no sign change or change in statistical significance for all other variables
. 
. * ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
. * Further testing for omitted variable bias in Multiple Imputation Model 4
. * ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
. * Under 5 child mortality--under5mortl
. * On 10 imputed datasets, testing Model 4 from Table 3.1
. clear

. use revspredictbycntryyr, clear

. drop if indstate==0
(6,647 observations deleted)

. misstable sum under5mortl, gen(miss)
                                                               Obs<.
                                                +------------------------------
               |                                | Unique
      Variable |     Obs=.     Obs>.     Obs<.  | values        Min         Max
  -------------+--------------------------------+------------------------------
   under5mortl |       205              11,537  |   >500        2.1       577.9
  -----------------------------------------------------------------------------

. mi set wide

. mi register regular lnpopl postcoldwar cowcode year urbancivicny 

. mi register imputed polityl gdppcthl  yrsincleaderinpower v2x_execorr lnoill under5mortl

. mi xtset cowcode year
       panel variable:  cowcode (unbalanced)
        time variable:  year, 1900 to 2014, but with gaps
                delta:  1 unit

. mi impute chained (pmm, knn(10)) polityl gdppcthl yrsincleaderinpower v2x_execorr lnoill under5mortl = lnpopl po
> stcoldwar, add(10) rseed(1234) force dots chaindots

Conditional models:
    yrsincleader~r: pmm yrsincleaderinpower v2x_execorr lnoill under5mortl polityl gdppcthl lnpopl postcoldwar ,
                     knn(10)
       v2x_execorr: pmm v2x_execorr yrsincleaderinpower lnoill under5mortl polityl gdppcthl lnpopl postcoldwar ,
                     knn(10)
            lnoill: pmm lnoill yrsincleaderinpower v2x_execorr under5mortl polityl gdppcthl lnpopl postcoldwar ,
                     knn(10)
       under5mortl: pmm under5mortl yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl lnpopl postcoldwar ,
                     knn(10)
           polityl: pmm polityl yrsincleaderinpower v2x_execorr lnoill under5mortl gdppcthl lnpopl postcoldwar ,
                     knn(10)
          gdppcthl: pmm gdppcthl yrsincleaderinpower v2x_execorr lnoill under5mortl polityl lnpopl postcoldwar ,
                     knn(10)

Performing chained iterations:
  imputing m=1: burn-in 10 .......... done
  imputing m=2: burn-in 10 .......... done
  imputing m=3: burn-in 10 .......... done
  imputing m=4: burn-in 10 .......... done
  imputing m=5: burn-in 10 .......... done
  imputing m=6: burn-in 10 .......... done
  imputing m=7: burn-in 10 .......... done
  imputing m=8: burn-in 10 .......... done
  imputing m=9: burn-in 10 .......... done
  imputing m=10: burn-in 10 .......... done

Multivariate imputation                     Imputations =       10
Chained equations                                 added =       10
Imputed: m=1 through m=10                       updated =        0

Initialization: monotone                     Iterations =      100
                                                burn-in =       10

           polityl: predictive mean matching
          gdppcthl: predictive mean matching
    yrsincleader~r: predictive mean matching
       v2x_execorr: predictive mean matching
            lnoill: predictive mean matching
       under5mortl: predictive mean matching

------------------------------------------------------------------
                   |               Observations per m             
                   |----------------------------------------------
          Variable |   Complete   Incomplete   Imputed |     Total
-------------------+-----------------------------------+----------
           polityl |      11328          414       414 |     11742
          gdppcthl |      11042          700       700 |     11742
    yrsincleader~r |      11661           81        81 |     11742
       v2x_execorr |      11580          162       162 |     11742
            lnoill |      11560          182       182 |     11742
       under5mortl |      11537          205       205 |     11742
------------------------------------------------------------------
(complete + incomplete = total; imputed is the minimum across m
 of the number of filled-in observations.)

. qui mi xeq 1: twoway (kdensity under5mortl if missunder5mortl==0) || (kdensity under5mortl if missunder5mortl==1
> ) || (kdensity under5mortl), legend(label(1 "Observed") label(2 "Imputed") label(3 "Completed"))

. graph export Robustnesstestfiles\Logfiles\mod4impobsunder5mortl.pdf, replace
(file Robustnesstestfiles\Logfiles\mod4impobsunder5mortl.pdf written in PDF format)

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny lnpopl c.gdppcthl##c.gdppcthl
>  c.polityl##c.polityl yrsincleaderinpower v2x_execorr lnoill postcoldwar under5mortl, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0084
                                                Largest FMI       =     0.0311
DF adjustment:   Large sample                   DF:     min       =   9,408.81
                                                        avg       =   6.59e+15
                                                        max       =   7.91e+16
Model F test:       Equal FMI                   F(  10,879964.4)  =      10.09
Within VCE type:       Robust                   Prob > F          =     0.0000

                                      (Within VCE adjusted for 164 clusters in cowcode)
---------------------------------------------------------------------------------------
         urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------------+----------------------------------------------------------------
               lnpopl |    1.50557   .1367414     4.51   0.000     1.260061    1.798913
             gdppcthl |   1.575265   .2107111     3.40   0.001     1.211971    2.047459
                      |
c.gdppcthl#c.gdppcthl |    .975205   .0077071    -3.18   0.001     .9602155    .9904285
                      |
              polityl |   .9213349   .0302926    -2.49   0.013      .863835    .9826622
                      |
  c.polityl#c.polityl |   .9842567    .005743    -2.72   0.007     .9730636    .9955786
                      |
  yrsincleaderinpower |   1.031881   .0130949     2.47   0.013     1.006532    1.057869
          v2x_execorr |    5.55415   4.720105     2.02   0.044     1.050117    29.37633
               lnoill |   .8800006    .026027    -4.32   0.000      .830439    .9325201
          postcoldwar |   4.821381   2.375851     3.19   0.001     1.835358     12.6655
          under5mortl |   .9953468    .002413    -1.92   0.054      .990628    1.000088
                _cons |   .0000163   .0000235    -7.63   0.000     9.55e-07    .0002766
----------------------+----------------------------------------------------------------
             /lnsig2u |  -9.599654   9245.853                     -18131.14    18111.94
----------------------+----------------------------------------------------------------
              sigma_u |   .0082312   38.05209                             0           .
                  rho |   .0000412   .3807902                             0           .
---------------------------------------------------------------------------------------

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny under5mortl time1 timesq time
> cub, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0027
                                                Largest FMI       =     0.0132
DF adjustment:   Large sample                   DF:     min       =  52,239.38
                                                        avg       =   3.97e+11
                                                        max       =   2.14e+12
Model F test:       Equal FMI                   F(   4, 1.9e+06)  =       7.86
Within VCE type:       Robust                   Prob > F          =     0.0000

                             (Within VCE adjusted for 164 clusters in cowcode)
------------------------------------------------------------------------------
urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
 under5mortl |   .9979424   .0018979    -1.08   0.279     .9942295    1.001669
       time1 |   .9425166   .0753212    -0.74   0.459     .8058705    1.102333
      timesq |   1.001581   .0013334     1.19   0.235     .9989706    1.004197
     timecub |   .9999924   6.69e-06    -1.14   0.254     .9999793    1.000005
       _cons |   .0009376   .0017936    -3.64   0.000     .0000221    .0398404
-------------+----------------------------------------------------------------
    /lnsig2u |  -1.514862   1.127029                     -3.723797    .6940734
-------------+----------------------------------------------------------------
     sigma_u |   .4688694   .2642146                      .1553773    1.414869
         rho |   .1178903   .1172021                      .0144644    .5489361
------------------------------------------------------------------------------

. *       RESULTS:  under5mortl--negative and marginally signficiant at the .10 level, did not change significance
>  or signs of other variables 
. *                 no bivariate relationship
. 
. * Percent urban--percurbanl
. * On 10 imputed datasets, testing Model 4 from Table 3.1
. clear

. use revspredictbycntryyr, clear

. drop if indstate==0
(6,647 observations deleted)

. misstable sum percurbanl, gen(miss)
                                                               Obs<.
                                                +------------------------------
               |                                | Unique
      Variable |     Obs=.     Obs>.     Obs<.  | values        Min         Max
  -------------+--------------------------------+------------------------------
    percurbanl |       424              11,318  |   >500   1.533333      99.061
  -----------------------------------------------------------------------------

. mi set wide

. mi register regular lnpopl postcoldwar cowcode year urbancivicny 

. mi register imputed polityl gdppcthl  yrsincleaderinpower v2x_execorr lnoill percurbanl

. mi xtset cowcode year
       panel variable:  cowcode (unbalanced)
        time variable:  year, 1900 to 2014, but with gaps
                delta:  1 unit

. mi impute chained (pmm, knn(10)) polityl gdppcthl yrsincleaderinpower v2x_execorr lnoill percurbanl = lnpopl pos
> tcoldwar, add(10) rseed(1234) force dots chaindots

Conditional models:
    yrsincleader~r: pmm yrsincleaderinpower v2x_execorr lnoill polityl percurbanl gdppcthl lnpopl postcoldwar ,
                     knn(10)
       v2x_execorr: pmm v2x_execorr yrsincleaderinpower lnoill polityl percurbanl gdppcthl lnpopl postcoldwar ,
                     knn(10)
            lnoill: pmm lnoill yrsincleaderinpower v2x_execorr polityl percurbanl gdppcthl lnpopl postcoldwar ,
                     knn(10)
           polityl: pmm polityl yrsincleaderinpower v2x_execorr lnoill percurbanl gdppcthl lnpopl postcoldwar ,
                     knn(10)
        percurbanl: pmm percurbanl yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl lnpopl postcoldwar ,
                     knn(10)
          gdppcthl: pmm gdppcthl yrsincleaderinpower v2x_execorr lnoill polityl percurbanl lnpopl postcoldwar ,
                     knn(10)

Performing chained iterations:
  imputing m=1: burn-in 10 .......... done
  imputing m=2: burn-in 10 .......... done
  imputing m=3: burn-in 10 .......... done
  imputing m=4: burn-in 10 .......... done
  imputing m=5: burn-in 10 .......... done
  imputing m=6: burn-in 10 .......... done
  imputing m=7: burn-in 10 .......... done
  imputing m=8: burn-in 10 .......... done
  imputing m=9: burn-in 10 .......... done
  imputing m=10: burn-in 10 .......... done

Multivariate imputation                     Imputations =       10
Chained equations                                 added =       10
Imputed: m=1 through m=10                       updated =        0

Initialization: monotone                     Iterations =      100
                                                burn-in =       10

           polityl: predictive mean matching
          gdppcthl: predictive mean matching
    yrsincleader~r: predictive mean matching
       v2x_execorr: predictive mean matching
            lnoill: predictive mean matching
        percurbanl: predictive mean matching

------------------------------------------------------------------
                   |               Observations per m             
                   |----------------------------------------------
          Variable |   Complete   Incomplete   Imputed |     Total
-------------------+-----------------------------------+----------
           polityl |      11328          414       414 |     11742
          gdppcthl |      11042          700       700 |     11742
    yrsincleader~r |      11661           81        81 |     11742
       v2x_execorr |      11580          162       162 |     11742
            lnoill |      11560          182       182 |     11742
        percurbanl |      11318          424       424 |     11742
------------------------------------------------------------------
(complete + incomplete = total; imputed is the minimum across m
 of the number of filled-in observations.)

. qui mi xeq 1: twoway (kdensity percurbanl  if misspercurbanl==0) || (kdensity percurbanl if misspercurbanl==1) |
> | (kdensity percurbanl), legend(label(1 "Observed") label(2 "Imputed") label(3 "Completed"))

. graph export Robustnesstestfiles\Logfiles\mod4impobspercurbanl.pdf, replace
(file Robustnesstestfiles\Logfiles\mod4impobspercurbanl.pdf written in PDF format)

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny lnpopl c.gdppcthl##c.gdppcthl
>  c.polityl##c.polityl yrsincleaderinpower v2x_execorr lnoill postcoldwar percurbanl, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0040
                                                Largest FMI       =     0.0256
DF adjustment:   Large sample                   DF:     min       =  13,843.21
                                                        avg       =   2.75e+20
                                                        max       =   3.30e+21
Model F test:       Equal FMI                   F(  10, 4.0e+06)  =      10.65
Within VCE type:       Robust                   Prob > F          =     0.0000

                                      (Within VCE adjusted for 164 clusters in cowcode)
---------------------------------------------------------------------------------------
         urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------------+----------------------------------------------------------------
               lnpopl |   1.528524   .1424377     4.55   0.000     1.273362    1.834817
             gdppcthl |    1.84581   .2775717     4.08   0.000     1.374623    2.478507
                      |
c.gdppcthl#c.gdppcthl |   .9665431   .0096338    -3.41   0.001     .9478445    .9856106
                      |
              polityl |   .9263343   .0301467    -2.35   0.019     .8690929    .9873459
                      |
  c.polityl#c.polityl |   .9846686   .0054772    -2.78   0.005     .9739908    .9954634
                      |
  yrsincleaderinpower |   1.034863   .0123127     2.88   0.004     1.011009    1.059279
          v2x_execorr |   5.387003   4.607703     1.97   0.049     1.007588    28.80125
               lnoill |   .8761542   .0254238    -4.56   0.000      .827715    .9274283
          postcoldwar |   7.410517   3.248223     4.57   0.000     3.138663    17.49655
           percurbanl |   1.000335    .010026     0.03   0.973     .9808764     1.02018
                _cons |   4.44e-06   6.09e-06    -8.97   0.000     3.00e-07    .0000655
----------------------+----------------------------------------------------------------
             /lnsig2u |  -10.06206   8982.548                     -17615.53    17595.41
----------------------+----------------------------------------------------------------
              sigma_u |   .0065321   29.33731                             0           .
                  rho |   .0000259   .2329861                             0           .
---------------------------------------------------------------------------------------

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny percurbanl time1 timesq timec
> ub, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0000
                                                Largest FMI       =     0.0000
DF adjustment:   Large sample                   DF:     min       =   5.34e+11
                                                        avg       =   1.52e+14
                                                        max       =   9.04e+14
Model F test:       Equal FMI                   F(   4, 1.3e+13)  =       6.48
Within VCE type:       Robust                   Prob > F          =     0.0000

                             (Within VCE adjusted for 164 clusters in cowcode)
------------------------------------------------------------------------------
urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
  percurbanl |   .9942739   .0060803    -0.94   0.348     .9824278    1.006263
       time1 |   .9468013   .0751918    -0.69   0.491     .8103249    1.106263
      timesq |   1.001644   .0013364     1.23   0.218     .9990285    1.004267
     timecub |    .999992   6.70e-06    -1.19   0.233     .9999789    1.000005
       _cons |   .0005471   .0009319    -4.41   0.000     .0000194     .015418
-------------+----------------------------------------------------------------
    /lnsig2u |  -1.790042   1.485225                     -4.701031    1.120946
-------------+----------------------------------------------------------------
     sigma_u |   .4085989   .3034307                        .09532    1.751501
         rho |   .0921432   .1242433                      .0054932    .6509563
------------------------------------------------------------------------------

. *       RESULTS:  percurbanl--positive and statistically insignificant, did not change significance or signs of 
> other variables
. *                 no bivariate relationship
. 
. * Average years of schooling--totalyrsschooll
. * On 10 imputed datasets, testing Model 4 from Table 3.1
. clear

. use revspredictbycntryyr, clear

. drop if indstate==0
(6,647 observations deleted)

. misstable sum totalyrsschooll, gen(miss)
                                                               Obs<.
                                                +------------------------------
               |                                | Unique
      Variable |     Obs=.     Obs>.     Obs<.  | values        Min         Max
  -------------+--------------------------------+------------------------------
  totalyrss~ll |     4,610               7,132  |   >500          0    13.42363
  -----------------------------------------------------------------------------

. mi set wide

. mi register regular lnpopl postcoldwar cowcode year urbancivicny 

. mi register imputed polityl gdppcthl  yrsincleaderinpower v2x_execorr lnoill totalyrsschooll

. mi xtset cowcode year
       panel variable:  cowcode (unbalanced)
        time variable:  year, 1900 to 2014, but with gaps
                delta:  1 unit

. mi impute chained (pmm, knn(10)) polityl gdppcthl yrsincleaderinpower v2x_execorr lnoill totalyrsschooll = lnpop
> l postcoldwar, add(10) rseed(1234) force dots chaindots

Conditional models:
    yrsincleader~r: pmm yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl totalyrsschooll lnpopl
                     postcoldwar , knn(10)
       v2x_execorr: pmm v2x_execorr yrsincleaderinpower lnoill polityl gdppcthl totalyrsschooll lnpopl
                     postcoldwar , knn(10)
            lnoill: pmm lnoill yrsincleaderinpower v2x_execorr polityl gdppcthl totalyrsschooll lnpopl
                     postcoldwar , knn(10)
           polityl: pmm polityl yrsincleaderinpower v2x_execorr lnoill gdppcthl totalyrsschooll lnpopl
                     postcoldwar , knn(10)
          gdppcthl: pmm gdppcthl yrsincleaderinpower v2x_execorr lnoill polityl totalyrsschooll lnpopl
                     postcoldwar , knn(10)
    totalyrssch~ll: pmm totalyrsschooll yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl lnpopl
                     postcoldwar , knn(10)

Performing chained iterations:
  imputing m=1: burn-in 10 .......... done
  imputing m=2: burn-in 10 .......... done
  imputing m=3: burn-in 10 .......... done
  imputing m=4: burn-in 10 .......... done
  imputing m=5: burn-in 10 .......... done
  imputing m=6: burn-in 10 .......... done
  imputing m=7: burn-in 10 .......... done
  imputing m=8: burn-in 10 .......... done
  imputing m=9: burn-in 10 .......... done
  imputing m=10: burn-in 10 .......... done

Multivariate imputation                     Imputations =       10
Chained equations                                 added =       10
Imputed: m=1 through m=10                       updated =        0

Initialization: monotone                     Iterations =      100
                                                burn-in =       10

           polityl: predictive mean matching
          gdppcthl: predictive mean matching
    yrsincleader~r: predictive mean matching
       v2x_execorr: predictive mean matching
            lnoill: predictive mean matching
    totalyrssch~ll: predictive mean matching

------------------------------------------------------------------
                   |               Observations per m             
                   |----------------------------------------------
          Variable |   Complete   Incomplete   Imputed |     Total
-------------------+-----------------------------------+----------
           polityl |      11328          414       414 |     11742
          gdppcthl |      11042          700       700 |     11742
    yrsincleader~r |      11661           81        81 |     11742
       v2x_execorr |      11580          162       162 |     11742
            lnoill |      11560          182       182 |     11742
    totalyrssch~ll |       7132         4610      4610 |     11742
------------------------------------------------------------------
(complete + incomplete = total; imputed is the minimum across m
 of the number of filled-in observations.)

. qui mi xeq 1: twoway (kdensity totalyrsschooll if misstotalyrsschooll==0) || (kdensity totalyrsschooll if missto
> talyrsschooll==1) || (kdensity totalyrsschooll), legend(label(1 "Observed") label(2 "Imputed") label(3 "Complete
> d"))

. graph export Robustnesstestfiles\Logfiles\mod4impobstotalyrsschooll.pdf, replace
(file Robustnesstestfiles\Logfiles\mod4impobstotalyrsschooll.pdf written in PDF format)

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny lnpopl c.gdppcthl##c.gdppcthl
>  c.polityl##c.polityl yrsincleaderinpower v2x_execorr lnoill postcoldwar totalyrsschooll, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0446
                                                Largest FMI       =     0.3122
DF adjustment:   Large sample                   DF:     min       =     100.84
                                                        avg       =   1.78e+15
                                                        max       =   2.14e+16
Model F test:       Equal FMI                   F(  10,34181.7)   =      10.01
Within VCE type:       Robust                   Prob > F          =     0.0000

                                      (Within VCE adjusted for 164 clusters in cowcode)
---------------------------------------------------------------------------------------
         urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------------+----------------------------------------------------------------
               lnpopl |   1.528589   .1384029     4.69   0.000     1.280031    1.825412
             gdppcthl |   1.807232   .2599954     4.11   0.000     1.362953    2.396331
                      |
c.gdppcthl#c.gdppcthl |   .9674209   .0090008    -3.56   0.000      .949938    .9852255
                      |
              polityl |   .9231045   .0320233    -2.31   0.021     .8624201    .9880589
                      |
  c.polityl#c.polityl |   .9850328   .0055392    -2.68   0.007     .9742349    .9959505
                      |
  yrsincleaderinpower |   1.034225   .0123173     2.83   0.005     1.010362     1.05865
          v2x_execorr |   5.365301   4.464485     2.02   0.043     1.050283    27.40829
               lnoill |   .8768288    .026024    -4.43   0.000     .8272779    .9293475
          postcoldwar |   7.288228   3.251317     4.45   0.000     3.039741     17.4746
      totalyrsschooll |   1.023622   .0835888     0.29   0.776     .8705354    1.203629
                _cons |   4.21e-06   5.43e-06    -9.59   0.000     3.35e-07    .0000528
----------------------+----------------------------------------------------------------
             /lnsig2u |  -9.642299   11198.99                     -21959.27    21939.98
----------------------+----------------------------------------------------------------
              sigma_u |   .0080575   45.11805                             0           .
                  rho |   .0000395   .4419762                             0           .
---------------------------------------------------------------------------------------

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny totalyrsschooll time1 timesq 
> timecub, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0475
                                                Largest FMI       =     0.2228
DF adjustment:   Large sample                   DF:     min       =     194.93
                                                        avg       = 4260241.93
                                                        max       =   1.50e+07
Model F test:       Equal FMI                   F(   4, 6781.5)   =       5.88
Within VCE type:       Robust                   Prob > F          =     0.0001

                                (Within VCE adjusted for 164 clusters in cowcode)
---------------------------------------------------------------------------------
   urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
totalyrsschooll |   .9490958   .0468913    -1.06   0.292     .8609791    1.046231
          time1 |   .9401164   .0742814    -0.78   0.434     .8052405    1.097584
         timesq |   1.001737   .0013403     1.30   0.195     .9991134    1.004367
        timecub |   .9999917   6.72e-06    -1.23   0.218     .9999786    1.000005
          _cons |   .0006015   .0010006    -4.46   0.000     .0000231    .0156713
----------------+----------------------------------------------------------------
       /lnsig2u |  -1.875437   1.656471                     -5.122062    1.371187
----------------+----------------------------------------------------------------
        sigma_u |     .39152   .3242708                      .0772251     1.98495
            rho |   .0852441   .1291676                      .0036124    .7054708
---------------------------------------------------------------------------------

. *       RESULTS:  totalyrsschooll--positive and statistically insignificant, did not change significance or sign
> s of other variables
. *                 bivariate relationship is negative and insignificant
. 
. * Military personnel per thousand pop--milperthousl
. * On 10 imputed datasets, testing Model 4 from Table 3.1
. clear

. use revspredictbycntryyr, clear

. drop if indstate==0
(6,647 observations deleted)

. misstable sum milperthousl, gen(miss)
                                                               Obs<.
                                                +------------------------------
               |                                | Unique
      Variable |     Obs=.     Obs>.     Obs<.  | values        Min         Max
  -------------+--------------------------------+------------------------------
  milperthousl |       704              11,038  |   >500          0    556.7766
  -----------------------------------------------------------------------------

. mi set wide

. mi register regular lnpopl postcoldwar cowcode year urbancivicny 

. mi register imputed polityl gdppcthl  yrsincleaderinpower v2x_execorr lnoill milperthousl

. mi xtset cowcode year
       panel variable:  cowcode (unbalanced)
        time variable:  year, 1900 to 2014, but with gaps
                delta:  1 unit

. mi impute chained (pmm, knn(10)) polityl gdppcthl yrsincleaderinpower v2x_execorr lnoill milperthousl = lnpopl p
> ostcoldwar, add(10) rseed(1234) force dots chaindots

Conditional models:
    yrsincleader~r: pmm yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl milperthousl lnpopl postcoldwar
                     , knn(10)
       v2x_execorr: pmm v2x_execorr yrsincleaderinpower lnoill polityl gdppcthl milperthousl lnpopl postcoldwar
                     , knn(10)
            lnoill: pmm lnoill yrsincleaderinpower v2x_execorr polityl gdppcthl milperthousl lnpopl postcoldwar
                     , knn(10)
           polityl: pmm polityl yrsincleaderinpower v2x_execorr lnoill gdppcthl milperthousl lnpopl postcoldwar
                     , knn(10)
          gdppcthl: pmm gdppcthl yrsincleaderinpower v2x_execorr lnoill polityl milperthousl lnpopl postcoldwar
                     , knn(10)
      milperthousl: pmm milperthousl yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl lnpopl postcoldwar
                     , knn(10)

Performing chained iterations:
  imputing m=1: burn-in 10 .......... done
  imputing m=2: burn-in 10 .......... done
  imputing m=3: burn-in 10 .......... done
  imputing m=4: burn-in 10 .......... done
  imputing m=5: burn-in 10 .......... done
  imputing m=6: burn-in 10 .......... done
  imputing m=7: burn-in 10 .......... done
  imputing m=8: burn-in 10 .......... done
  imputing m=9: burn-in 10 .......... done
  imputing m=10: burn-in 10 .......... done

Multivariate imputation                     Imputations =       10
Chained equations                                 added =       10
Imputed: m=1 through m=10                       updated =        0

Initialization: monotone                     Iterations =      100
                                                burn-in =       10

           polityl: predictive mean matching
          gdppcthl: predictive mean matching
    yrsincleader~r: predictive mean matching
       v2x_execorr: predictive mean matching
            lnoill: predictive mean matching
      milperthousl: predictive mean matching

------------------------------------------------------------------
                   |               Observations per m             
                   |----------------------------------------------
          Variable |   Complete   Incomplete   Imputed |     Total
-------------------+-----------------------------------+----------
           polityl |      11328          414       414 |     11742
          gdppcthl |      11042          700       700 |     11742
    yrsincleader~r |      11661           81        81 |     11742
       v2x_execorr |      11580          162       162 |     11742
            lnoill |      11560          182       182 |     11742
      milperthousl |      11038          704       704 |     11742
------------------------------------------------------------------
(complete + incomplete = total; imputed is the minimum across m
 of the number of filled-in observations.)

. qui mi xeq 1: twoway (kdensity milperthousl if missmilperthousl==0) || (kdensity milperthousl if missmilperthous
> l==1) || (kdensity milperthousl), legend(label(1 "Observed") label(2 "Imputed") label(3 "Completed"))

. graph export Robustnesstestfiles\Logfiles\mod4impobsmilperthousl.pdf, replace
(file Robustnesstestfiles\Logfiles\mod4impobsmilperthousl.pdf written in PDF format)

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny lnpopl c.gdppcthl##c.gdppcthl
>  c.polityl##c.polityl yrsincleaderinpower v2x_execorr lnoill postcoldwar milperthousl, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0082
                                                Largest FMI       =     0.0406
DF adjustment:   Large sample                   DF:     min       =   5,550.36
                                                        avg       =   9.29e+18
                                                        max       =   1.12e+20
Model F test:       Equal FMI                   F(  10,893903.7)  =      10.87
Within VCE type:       Robust                   Prob > F          =     0.0000

                                      (Within VCE adjusted for 164 clusters in cowcode)
---------------------------------------------------------------------------------------
         urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------------+----------------------------------------------------------------
               lnpopl |   1.522706   .1356224     4.72   0.000       1.2788    1.813134
             gdppcthl |   1.956082   .2619204     5.01   0.000     1.504562    2.543102
                      |
c.gdppcthl#c.gdppcthl |   .9632304   .0093366    -3.86   0.000     .9451036    .9817048
                      |
              polityl |   .9209285   .0304117    -2.49   0.013     .8632045    .9825125
                      |
  c.polityl#c.polityl |   .9860477   .0055104    -2.51   0.012      .975304    .9969097
                      |
  yrsincleaderinpower |   1.039704   .0125912     3.22   0.001     1.015316    1.064678
          v2x_execorr |   5.172321   4.436777     1.92   0.055     .9627889    27.78689
               lnoill |   .8685976   .0264038    -4.63   0.000     .8183584    .9219209
          postcoldwar |   7.629613   3.238958     4.79   0.000     3.320064    17.53309
         milperthousl |   .9756177   .0161007    -1.50   0.135     .9445653    1.007691
                _cons |   4.58e-06   5.76e-06    -9.77   0.000     3.89e-07     .000054
----------------------+----------------------------------------------------------------
             /lnsig2u |  -11.28964   43308.19                     -84893.78     84871.2
----------------------+----------------------------------------------------------------
              sigma_u |   .0035358   76.56425                             0           .
                  rho |   7.60e-06   .3291448                             0           .
---------------------------------------------------------------------------------------

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny milperthousl time1 timesq tim
> ecub, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0013
                                                Largest FMI       =     0.0073
DF adjustment:   Large sample                   DF:     min       = 169,737.59
                                                        avg       =   4.96e+10
                                                        max       =   1.26e+11
Model F test:       Equal FMI                   F(   4, 7.8e+06)  =       6.17
Within VCE type:       Robust                   Prob > F          =     0.0001

                             (Within VCE adjusted for 164 clusters in cowcode)
------------------------------------------------------------------------------
urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
milperthousl |   1.006053   .0071629     0.85   0.397     .9921112     1.02019
       time1 |   .9481406   .0766298    -0.66   0.510     .8092406    1.110882
      timesq |     1.0016   .0013385     1.20   0.232       .99898    1.004227
     timecub |   .9999922   6.69e-06    -1.17   0.243     .9999791    1.000005
       _cons |   .0004323   .0007586    -4.41   0.000     .0000139    .0134778
-------------+----------------------------------------------------------------
    /lnsig2u |  -1.662011   1.304071                     -4.217943    .8939202
-------------+----------------------------------------------------------------
     sigma_u |    .435611   .2840338                      .1213627    1.563552
         rho |   .1034272   .1209265                      .0088746    .5977791
------------------------------------------------------------------------------

. *       RESULTS:  milperthousl--negative and statistically insignificant, did not change significance or signs o
> f other variables
. *                 no bivariate relationship
. 
. * Population density--lnpopdensityl
. * Does not need multiple imputation, testing Model 4 from Table 3.1
. clear

. use revspredictbycntryyr, clear

. drop if indstate==0
(6,647 observations deleted)

. mi set wide

. mi register regular lnpopl postcoldwar cowcode year urbancivicny lnpopdensityl

. mi register imputed polityl gdppcthl  yrsincleaderinpower v2x_execorr lnoill 

. mi xtset cowcode year
       panel variable:  cowcode (unbalanced)
        time variable:  year, 1900 to 2014, but with gaps
                delta:  1 unit

. mi impute chained (pmm, knn(10)) polityl gdppcthl yrsincleaderinpower v2x_execorr lnoill = lnpopl postcoldwar ln
> popdensityl, add(10) rseed(1234) force dots chaindots

Conditional models:
    yrsincleader~r: pmm yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl lnpopl postcoldwar lnpopdensityl
                     , knn(10)
       v2x_execorr: pmm v2x_execorr yrsincleaderinpower lnoill polityl gdppcthl lnpopl postcoldwar lnpopdensityl
                     , knn(10)
            lnoill: pmm lnoill yrsincleaderinpower v2x_execorr polityl gdppcthl lnpopl postcoldwar lnpopdensityl
                     , knn(10)
           polityl: pmm polityl yrsincleaderinpower v2x_execorr lnoill gdppcthl lnpopl postcoldwar lnpopdensityl
                     , knn(10)
          gdppcthl: pmm gdppcthl yrsincleaderinpower v2x_execorr lnoill polityl lnpopl postcoldwar lnpopdensityl
                     , knn(10)

Performing chained iterations:
  imputing m=1: burn-in 10 .......... done
  imputing m=2: burn-in 10 .......... done
  imputing m=3: burn-in 10 .......... done
  imputing m=4: burn-in 10 .......... done
  imputing m=5: burn-in 10 .......... done
  imputing m=6: burn-in 10 .......... done
  imputing m=7: burn-in 10 .......... done
  imputing m=8: burn-in 10 .......... done
  imputing m=9: burn-in 10 .......... done
  imputing m=10: burn-in 10 .......... done

Multivariate imputation                     Imputations =       10
Chained equations                                 added =       10
Imputed: m=1 through m=10                       updated =        0

Initialization: monotone                     Iterations =      100
                                                burn-in =       10

           polityl: predictive mean matching
          gdppcthl: predictive mean matching
    yrsincleader~r: predictive mean matching
       v2x_execorr: predictive mean matching
            lnoill: predictive mean matching

------------------------------------------------------------------
                   |               Observations per m             
                   |----------------------------------------------
          Variable |   Complete   Incomplete   Imputed |     Total
-------------------+-----------------------------------+----------
           polityl |      11328          414       414 |     11742
          gdppcthl |      11042          700       700 |     11742
    yrsincleader~r |      11661           81        81 |     11742
       v2x_execorr |      11580          162       162 |     11742
            lnoill |      11560          182       182 |     11742
------------------------------------------------------------------
(complete + incomplete = total; imputed is the minimum across m
 of the number of filled-in observations.)

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny lnpopl c.gdppcthl##c.gdppcthl
>  c.polityl##c.polityl yrsincleaderinpower v2x_execorr lnoill postcoldwar lnpopdensityl, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0047
                                                Largest FMI       =     0.0235
DF adjustment:   Large sample                   DF:     min       =  16,453.44
                                                        avg       =   1.88e+19
                                                        max       =   2.25e+20
Model F test:       Equal FMI                   F(  10, 3.3e+06)  =      10.37
Within VCE type:       Robust                   Prob > F          =     0.0000

                                      (Within VCE adjusted for 164 clusters in cowcode)
---------------------------------------------------------------------------------------
         urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------------+----------------------------------------------------------------
               lnpopl |   1.453554   .1254201     4.33   0.000     1.227397    1.721381
             gdppcthl |   1.760876    .209448     4.76   0.000     1.394705    2.223183
                      |
c.gdppcthl#c.gdppcthl |   .9692257   .0079187    -3.83   0.000      .953829     .984871
                      |
              polityl |   .9262116   .0300278    -2.36   0.018     .8691889    .9869752
                      |
  c.polityl#c.polityl |   .9848458   .0057215    -2.63   0.009     .9736946    .9961247
                      |
  yrsincleaderinpower |   1.034893     .01238     2.87   0.004     1.010911    1.059444
          v2x_execorr |   5.311932   4.487061     1.98   0.048     1.014451    27.81468
               lnoill |    .894321   .0290834    -3.43   0.001     .8390972    .9531793
          postcoldwar |   6.888919   3.055825     4.35   0.000     2.887822    16.43357
        lnpopdensityl |   1.218547   .1650106     1.46   0.144     .9344934    1.588944
                _cons |   3.60e-06   4.77e-06    -9.46   0.000     2.68e-07    .0000483
----------------------+----------------------------------------------------------------
             /lnsig2u |  -11.35765   46463.12                     -91077.41    91054.69
----------------------+----------------------------------------------------------------
              sigma_u |   .0034176    79.3955                             0           .
                  rho |   7.10e-06   .3299048                             0           .
---------------------------------------------------------------------------------------

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny lnpopdensityl time1 timesq ti
> mecub, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0000
                                                Largest FMI       =     0.0000
DF adjustment:   Large sample                   DF:     min       =   1.13e+63
                                                        avg       =   3.81e+64
                                                        max       =          .
Model F test:       Equal FMI                   F(   4, 1.9e+60)  =       8.34
Within VCE type:       Robust                   Prob > F          =     0.0000

                              (Within VCE adjusted for 164 clusters in cowcode)
-------------------------------------------------------------------------------
 urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
--------------+----------------------------------------------------------------
lnpopdensityl |   1.257399   .1290367     2.23   0.026     1.028303    1.537535
        time1 |   .9468991    .075459    -0.68   0.494     .8099736    1.106972
       timesq |   1.001579   .0013267     1.19   0.234     .9989821    1.004183
      timecub |   .9999922   6.65e-06    -1.17   0.244     .9999792    1.000005
        _cons |   .0002425   .0004099    -4.92   0.000     8.83e-06     .006661
--------------+----------------------------------------------------------------
     /lnsig2u |  -1.954086   1.689017                     -5.264498    1.356326
--------------+----------------------------------------------------------------
      sigma_u |   .3764225    .317892                      .0719165    1.970255
          rho |    .079308    .123329                      .0031343    .7023734
-------------------------------------------------------------------------------

. *       RESULTS:  lnpopdensityl--positive but statistically insignificant; did not change significance or signs 
> of other variables
. *                 bivariate relationship significant at the .05 level, controlling for time
. 
. * Youth bulge--youthpercl
. * On 10 imputed datasets, testing Model 4 from Table 3.1
. clear

. use revspredictbycntryyr, clear

. drop if indstate==0
(6,647 observations deleted)

. misstable sum youthpercl, gen(miss)
                                                               Obs<.
                                                +------------------------------
               |                                | Unique
      Variable |     Obs=.     Obs>.     Obs<.  | values        Min         Max
  -------------+--------------------------------+------------------------------
    youthpercl |     3,215               8,527  |   >500   11.22343    42.06414
  -----------------------------------------------------------------------------

. mi set wide

. mi register regular lnpopl postcoldwar cowcode year urbancivicny 

. mi register imputed polityl gdppcthl  yrsincleaderinpower v2x_execorr lnoill youthpercl

. mi xtset cowcode year
       panel variable:  cowcode (unbalanced)
        time variable:  year, 1900 to 2014, but with gaps
                delta:  1 unit

. mi impute chained (pmm, knn(10)) polityl gdppcthl yrsincleaderinpower v2x_execorr lnoill youthpercl = lnpopl pos
> tcoldwar, add(10) rseed(1234) force dots chaindots

Conditional models:
    yrsincleader~r: pmm yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl youthpercl lnpopl postcoldwar ,
                     knn(10)
       v2x_execorr: pmm v2x_execorr yrsincleaderinpower lnoill polityl gdppcthl youthpercl lnpopl postcoldwar ,
                     knn(10)
            lnoill: pmm lnoill yrsincleaderinpower v2x_execorr polityl gdppcthl youthpercl lnpopl postcoldwar ,
                     knn(10)
           polityl: pmm polityl yrsincleaderinpower v2x_execorr lnoill gdppcthl youthpercl lnpopl postcoldwar ,
                     knn(10)
          gdppcthl: pmm gdppcthl yrsincleaderinpower v2x_execorr lnoill polityl youthpercl lnpopl postcoldwar ,
                     knn(10)
        youthpercl: pmm youthpercl yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl lnpopl postcoldwar ,
                     knn(10)

Performing chained iterations:
  imputing m=1: burn-in 10 .......... done
  imputing m=2: burn-in 10 .......... done
  imputing m=3: burn-in 10 .......... done
  imputing m=4: burn-in 10 .......... done
  imputing m=5: burn-in 10 .......... done
  imputing m=6: burn-in 10 .......... done
  imputing m=7: burn-in 10 .......... done
  imputing m=8: burn-in 10 .......... done
  imputing m=9: burn-in 10 .......... done
  imputing m=10: burn-in 10 .......... done

Multivariate imputation                     Imputations =       10
Chained equations                                 added =       10
Imputed: m=1 through m=10                       updated =        0

Initialization: monotone                     Iterations =      100
                                                burn-in =       10

           polityl: predictive mean matching
          gdppcthl: predictive mean matching
    yrsincleader~r: predictive mean matching
       v2x_execorr: predictive mean matching
            lnoill: predictive mean matching
        youthpercl: predictive mean matching

------------------------------------------------------------------
                   |               Observations per m             
                   |----------------------------------------------
          Variable |   Complete   Incomplete   Imputed |     Total
-------------------+-----------------------------------+----------
           polityl |      11328          414       414 |     11742
          gdppcthl |      11042          700       700 |     11742
    yrsincleader~r |      11661           81        81 |     11742
       v2x_execorr |      11580          162       162 |     11742
            lnoill |      11560          182       182 |     11742
        youthpercl |       8527         3215      3215 |     11742
------------------------------------------------------------------
(complete + incomplete = total; imputed is the minimum across m
 of the number of filled-in observations.)

. qui mi xeq 1: twoway (kdensity youthpercl if missyouthpercl==0) || (kdensity youthpercl if missyouthpercl==1) ||
>  (kdensity youthpercl), legend(label(1 "Observed") label(2 "Imputed") label(3 "Completed"))

. graph export Robustnesstestfiles\Logfiles\mod4impobsyouthpercl.pdf, replace
(file Robustnesstestfiles\Logfiles\mod4impobsyouthpercl.pdf written in PDF format)

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny lnpopl c.gdppcthl##c.gdppcthl
>  c.polityl##c.polityl yrsincleaderinpower v2x_execorr lnoill postcoldwar youthpercl, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0065
                                                Largest FMI       =     0.0396
DF adjustment:   Large sample                   DF:     min       =   5,844.79
                                                        avg       =   3.93e+18
                                                        max       =   4.71e+19
Model F test:       Equal FMI                   F(  10, 1.5e+06)  =      10.76
Within VCE type:       Robust                   Prob > F          =     0.0000

                                      (Within VCE adjusted for 164 clusters in cowcode)
---------------------------------------------------------------------------------------
         urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------------+----------------------------------------------------------------
               lnpopl |   1.500602   .1346995     4.52   0.000     1.258516    1.789256
             gdppcthl |   1.697755   .2387274     3.76   0.000     1.288792    2.236492
                      |
c.gdppcthl#c.gdppcthl |   .9696882   .0086678    -3.44   0.001     .9528474    .9868266
                      |
              polityl |   .9239387    .030061    -2.43   0.015     .8668594    .9847764
                      |
  c.polityl#c.polityl |   .9841892   .0055999    -2.80   0.005      .973274    .9952269
                      |
  yrsincleaderinpower |   1.036077   .0123733     2.97   0.003     1.012108    1.060615
          v2x_execorr |   5.692451   4.857776     2.04   0.042     1.068828     30.3173
               lnoill |    .881152   .0256893    -4.34   0.000     .8322135    .9329685
          postcoldwar |   7.743649   3.392597     4.67   0.000     3.281124    18.27547
           youthpercl |   .9651235   .0269487    -1.27   0.204      .913714    1.019425
                _cons |   .0000179   .0000287    -6.81   0.000     7.68e-07    .0004168
----------------------+----------------------------------------------------------------
             /lnsig2u |   -10.3085   13658.39                     -26780.26    26759.64
----------------------+----------------------------------------------------------------
              sigma_u |   .0057748   39.43725                             0           .
                  rho |   .0000203   .2768903                             0           .
---------------------------------------------------------------------------------------

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny youthpercl time1 timesq timec
> ub, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0041
                                                Largest FMI       =     0.0239
DF adjustment:   Large sample                   DF:     min       =  15,973.55
                                                        avg       =   4.98e+08
                                                        max       =   1.19e+09
Model F test:       Equal FMI                   F(   4,771069.2)  =       6.30
Within VCE type:       Robust                   Prob > F          =     0.0000

                             (Within VCE adjusted for 164 clusters in cowcode)
------------------------------------------------------------------------------
urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
  youthpercl |   1.012206   .0177347     0.69   0.489     .9780343    1.047572
       time1 |   .9506661   .0762536    -0.63   0.528     .8123674    1.112509
      timesq |   1.001541   .0013344     1.16   0.248     .9989287    1.004159
     timecub |   .9999925   6.68e-06    -1.12   0.264     .9999795    1.000006
       _cons |   .0003099   .0005601    -4.47   0.000     8.97e-06     .010706
-------------+----------------------------------------------------------------
    /lnsig2u |  -1.669258   1.314674                     -4.245972    .9074556
-------------+----------------------------------------------------------------
     sigma_u |   .4340354   .2853076                      .1196737    1.574169
         rho |   .1027571   .1212104                      .0086315    .6010292
------------------------------------------------------------------------------

. *       RESULTS:  youthpercl--negative and statistically insignificant, did not change significance or signs of 
> other variables
. *                 positive and insignificant in the bivariate relationship
. 
. * Nominal trade as percent of nominal GDP--tottradepernomgdpl
. * On 10 imputed datasets, testing Model 4 from Table 3.1
. clear

. use revspredictbycntryyr, clear

. drop if indstate==0
(6,647 observations deleted)

. misstable sum tottradepernomgdpl, gen(miss)
                                                               Obs<.
                                                +------------------------------
               |                                | Unique
      Variable |     Obs=.     Obs>.     Obs<.  | values        Min         Max
  -------------+--------------------------------+------------------------------
  tottradepe~l |     2,432               9,310  |   >500   .0012177    51.00711
  -----------------------------------------------------------------------------

. mi set wide

. mi register regular lnpopl postcoldwar cowcode year urbancivicny 

. mi register imputed polityl gdppcthl  yrsincleaderinpower v2x_execorr lnoill tottradepernomgdpl

. mi xtset cowcode year
       panel variable:  cowcode (unbalanced)
        time variable:  year, 1900 to 2014, but with gaps
                delta:  1 unit

. mi impute chained (pmm, knn(10)) polityl gdppcthl yrsincleaderinpower v2x_execorr lnoill tottradepernomgdpl = ln
> popl postcoldwar, add(10) rseed(1234) force dots chaindots

Conditional models:
    yrsincleader~r: pmm yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl tottradepernomgdpl lnpopl
                     postcoldwar , knn(10)
       v2x_execorr: pmm v2x_execorr yrsincleaderinpower lnoill polityl gdppcthl tottradepernomgdpl lnpopl
                     postcoldwar , knn(10)
            lnoill: pmm lnoill yrsincleaderinpower v2x_execorr polityl gdppcthl tottradepernomgdpl lnpopl
                     postcoldwar , knn(10)
           polityl: pmm polityl yrsincleaderinpower v2x_execorr lnoill gdppcthl tottradepernomgdpl lnpopl
                     postcoldwar , knn(10)
          gdppcthl: pmm gdppcthl yrsincleaderinpower v2x_execorr lnoill polityl tottradepernomgdpl lnpopl
                     postcoldwar , knn(10)
    tottradepern~l: pmm tottradepernomgdpl yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl lnpopl
                     postcoldwar , knn(10)

Performing chained iterations:
  imputing m=1: burn-in 10 .......... done
  imputing m=2: burn-in 10 .......... done
  imputing m=3: burn-in 10 .......... done
  imputing m=4: burn-in 10 .......... done
  imputing m=5: burn-in 10 .......... done
  imputing m=6: burn-in 10 .......... done
  imputing m=7: burn-in 10 .......... done
  imputing m=8: burn-in 10 .......... done
  imputing m=9: burn-in 10 .......... done
  imputing m=10: burn-in 10 .......... done

Multivariate imputation                     Imputations =       10
Chained equations                                 added =       10
Imputed: m=1 through m=10                       updated =        0

Initialization: monotone                     Iterations =      100
                                                burn-in =       10

           polityl: predictive mean matching
          gdppcthl: predictive mean matching
    yrsincleader~r: predictive mean matching
       v2x_execorr: predictive mean matching
            lnoill: predictive mean matching
    tottradepern~l: predictive mean matching

------------------------------------------------------------------
                   |               Observations per m             
                   |----------------------------------------------
          Variable |   Complete   Incomplete   Imputed |     Total
-------------------+-----------------------------------+----------
           polityl |      11328          414       414 |     11742
          gdppcthl |      11042          700       700 |     11742
    yrsincleader~r |      11661           81        81 |     11742
       v2x_execorr |      11580          162       162 |     11742
            lnoill |      11560          182       182 |     11742
    tottradepern~l |       9310         2432      2432 |     11742
------------------------------------------------------------------
(complete + incomplete = total; imputed is the minimum across m
 of the number of filled-in observations.)

. qui mi xeq 1: twoway (kdensity tottradepernomgdpl if misstottradepernomgdpl==0) || (kdensity tottradepernomgdpl 
> if misstottradepernomgdpl==1) || (kdensity tottradepernomgdpl), legend(label(1 "Observed") label(2 "Imputed") la
> bel(3 "Completed"))

. graph export Robustnesstestfiles\Logfiles\mod4impobstottradepernomgdpl.pdf, replace
(file Robustnesstestfiles\Logfiles\mod4impobstottradepernomgdpl.pdf written in PDF format)

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny lnpopl c.gdppcthl##c.gdppcthl
>  c.polityl##c.polityl yrsincleaderinpower v2x_execorr lnoill postcoldwar tottradepernomgdpl, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0075
                                                Largest FMI       =     0.0577
DF adjustment:   Large sample                   DF:     min       =   2,766.32
                                                        avg       =   1.69e+20
                                                        max       =   2.02e+21
Model F test:       Equal FMI                   F(  10, 1.1e+06)  =      10.48
Within VCE type:       Robust                   Prob > F          =     0.0000

                                      (Within VCE adjusted for 164 clusters in cowcode)
---------------------------------------------------------------------------------------
         urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------------+----------------------------------------------------------------
               lnpopl |   1.524318   .1345996     4.77   0.000     1.282075    1.812333
             gdppcthl |   1.840327   .2380338     4.72   0.000     1.428229    2.371331
                      |
c.gdppcthl#c.gdppcthl |   .9667681   .0088964    -3.67   0.000     .9494879    .9843629
                      |
              polityl |   .9264629   .0301625    -2.35   0.019     .8691921    .9875071
                      |
  c.polityl#c.polityl |   .9846532   .0054036    -2.82   0.005     .9741188    .9953015
                      |
  yrsincleaderinpower |   1.034901   .0121961     2.91   0.004     1.011271    1.059083
          v2x_execorr |   5.292873   4.426763     1.99   0.046     1.027491    27.26497
               lnoill |   .8769088   .0257135    -4.48   0.000     .8279321    .9287828
          postcoldwar |   7.668494   3.297872     4.74   0.000      3.30101    17.81449
   tottradepernomgdpl |   .9687066   .1183437    -0.26   0.795     .7623567     1.23091
                _cons |   4.64e-06   5.84e-06    -9.77   0.000     3.95e-07    .0000545
----------------------+----------------------------------------------------------------
             /lnsig2u |  -10.01325   8359.556                     -16394.44    16374.41
----------------------+----------------------------------------------------------------
              sigma_u |   .0066934   27.97711                             0           .
                  rho |   .0000272   .2276723                             0           .
---------------------------------------------------------------------------------------

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny tottradepernomgdpl time1 time
> sq timecub, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0160
                                                Largest FMI       =     0.0861
DF adjustment:   Large sample                   DF:     min       =   1,257.46
                                                        avg       =   1.09e+08
                                                        max       =   3.76e+08
Model F test:       Equal FMI                   F(   4,54715.5)   =       6.08
Within VCE type:       Robust                   Prob > F          =     0.0001

                                   (Within VCE adjusted for 164 clusters in cowcode)
------------------------------------------------------------------------------------
      urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------------+----------------------------------------------------------------
tottradepernomgdpl |   .8510584   .2763748    -0.50   0.620     .4500614    1.609337
             time1 |   .9450542   .0751685    -0.71   0.477     .8086361    1.104486
            timesq |   1.001649   .0013306     1.24   0.215     .9990446    1.004261
           timecub |    .999992   6.66e-06    -1.20   0.230     .9999789    1.000005
             _cons |   .0005232   .0008888    -4.45   0.000     .0000187    .0146155
-------------------+----------------------------------------------------------------
          /lnsig2u |  -1.750014   1.384248                      -4.46309    .9630621
-------------------+----------------------------------------------------------------
           sigma_u |   .4168591   .2885182                      .1073624    1.618551
               rho |   .0955468   .1196234                      .0069586    .6142854
------------------------------------------------------------------------------------

. *       RESULTS:  tottradepernomgdpl--negative and statistically insignificant, did not change significance or s
> igns of other variables
. *                 no bivariate relationship
. 
. * Dollar exchange rate--lndollexchratel
. * On 10 imputed datasets, testing Model 4 from Table 3.1
. clear

. use revspredictbycntryyr, clear

. drop if indstate==0
(6,647 observations deleted)

. misstable sum lndollexchratel, gen(miss)
                                                               Obs<.
                                                +------------------------------
               |                                | Unique
      Variable |     Obs=.     Obs>.     Obs<.  | values        Min         Max
  -------------+--------------------------------+------------------------------
  lndollexch~l |     2,916               8,826  |   >500          0    28.91816
  -----------------------------------------------------------------------------

. mi set wide

. mi register regular lnpopl postcoldwar cowcode year urbancivicny 

. mi register imputed polityl gdppcthl  yrsincleaderinpower v2x_execorr lnoill lndollexchratel

. mi xtset cowcode year
       panel variable:  cowcode (unbalanced)
        time variable:  year, 1900 to 2014, but with gaps
                delta:  1 unit

. mi impute chained (pmm, knn(10)) polityl gdppcthl yrsincleaderinpower v2x_execorr lnoill lndollexchratel = lnpop
> l postcoldwar, add(10) rseed(1234) force dots chaindots

Conditional models:
    yrsincleader~r: pmm yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl lndollexchratel lnpopl
                     postcoldwar , knn(10)
       v2x_execorr: pmm v2x_execorr yrsincleaderinpower lnoill polityl gdppcthl lndollexchratel lnpopl
                     postcoldwar , knn(10)
            lnoill: pmm lnoill yrsincleaderinpower v2x_execorr polityl gdppcthl lndollexchratel lnpopl
                     postcoldwar , knn(10)
           polityl: pmm polityl yrsincleaderinpower v2x_execorr lnoill gdppcthl lndollexchratel lnpopl
                     postcoldwar , knn(10)
          gdppcthl: pmm gdppcthl yrsincleaderinpower v2x_execorr lnoill polityl lndollexchratel lnpopl
                     postcoldwar , knn(10)
    lndollexchra~l: pmm lndollexchratel yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl lnpopl
                     postcoldwar , knn(10)

Performing chained iterations:
  imputing m=1: burn-in 10 .......... done
  imputing m=2: burn-in 10 .......... done
  imputing m=3: burn-in 10 .......... done
  imputing m=4: burn-in 10 .......... done
  imputing m=5: burn-in 10 .......... done
  imputing m=6: burn-in 10 .......... done
  imputing m=7: burn-in 10 .......... done
  imputing m=8: burn-in 10 .......... done
  imputing m=9: burn-in 10 .......... done
  imputing m=10: burn-in 10 .......... done

Multivariate imputation                     Imputations =       10
Chained equations                                 added =       10
Imputed: m=1 through m=10                       updated =        0

Initialization: monotone                     Iterations =      100
                                                burn-in =       10

           polityl: predictive mean matching
          gdppcthl: predictive mean matching
    yrsincleader~r: predictive mean matching
       v2x_execorr: predictive mean matching
            lnoill: predictive mean matching
    lndollexchra~l: predictive mean matching

------------------------------------------------------------------
                   |               Observations per m             
                   |----------------------------------------------
          Variable |   Complete   Incomplete   Imputed |     Total
-------------------+-----------------------------------+----------
           polityl |      11328          414       414 |     11742
          gdppcthl |      11042          700       700 |     11742
    yrsincleader~r |      11661           81        81 |     11742
       v2x_execorr |      11580          162       162 |     11742
            lnoill |      11560          182       182 |     11742
    lndollexchra~l |       8826         2916      2916 |     11742
------------------------------------------------------------------
(complete + incomplete = total; imputed is the minimum across m
 of the number of filled-in observations.)

. qui mi xeq 1: twoway (kdensity lndollexchratel if misslndollexchratel==0) || (kdensity lndollexchratel if missln
> dollexchratel==1) || (kdensity lndollexchratel), legend(label(1 "Observed") label(2 "Imputed") label(3 "Complete
> d"))

. graph export Robustnesstestfiles\Logfiles\mod4impobslndollexchratel.pdf, replace
(file Robustnesstestfiles\Logfiles\mod4impobslndollexchratel.pdf written in PDF format)

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny lnpopl c.gdppcthl##c.gdppcthl
>  c.polityl##c.polityl yrsincleaderinpower v2x_execorr lnoill postcoldwar lndollexchratel, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0223
                                                Largest FMI       =     0.1686
DF adjustment:   Large sample                   DF:     min       =     335.87
                                                        avg       =   2.33e+18
                                                        max       =   2.80e+19
Model F test:       Equal FMI                   F(  10,125185.5)  =      10.25
Within VCE type:       Robust                   Prob > F          =     0.0000

                                      (Within VCE adjusted for 164 clusters in cowcode)
---------------------------------------------------------------------------------------
         urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------------+----------------------------------------------------------------
               lnpopl |   1.528321   .1383804     4.68   0.000     1.279803    1.825097
             gdppcthl |   1.842463   .2407634     4.68   0.000     1.426159    2.380289
                      |
c.gdppcthl#c.gdppcthl |   .9664899   .0091223    -3.61   0.000     .9487748    .9845357
                      |
              polityl |   .9266907   .0300039    -2.35   0.019     .8697107    .9874039
                      |
  c.polityl#c.polityl |   .9855795    .005548    -2.58   0.010     .9747645    .9965146
                      |
  yrsincleaderinpower |     1.0349   .0122958     2.89   0.004     1.011079    1.059282
          v2x_execorr |   5.238821    4.47908     1.94   0.053     .9805583    27.98941
               lnoill |   .8779404   .0274613    -4.16   0.000     .8257326    .9334492
          postcoldwar |   7.528239   3.303645     4.60   0.000     3.185336    17.79228
      lndollexchratel |   1.008887   .0650305     0.14   0.891     .8887463    1.145269
                _cons |   4.24e-06   5.41e-06    -9.68   0.000     3.46e-07    .0000518
----------------------+----------------------------------------------------------------
             /lnsig2u |  -10.15835   10086.39                     -19779.12     19758.8
----------------------+----------------------------------------------------------------
              sigma_u |   .0062251   31.39417                             0           .
                  rho |   .0000236   .2376038                             0           .
---------------------------------------------------------------------------------------

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny lndollexchratel time1 timesq 
> timecub, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0263
                                                Largest FMI       =     0.1383
DF adjustment:   Large sample                   DF:     min       =     495.15
                                                        avg       =   4.24e+07
                                                        max       =   1.71e+08
Model F test:       Equal FMI                   F(   4,19981.6)   =       6.55
Within VCE type:       Robust                   Prob > F          =     0.0000

                                (Within VCE adjusted for 164 clusters in cowcode)
---------------------------------------------------------------------------------
   urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
lndollexchratel |   1.068002   .0645819     1.09   0.277     .9483617    1.202736
          time1 |     .95282    .077671    -0.59   0.553     .8121261    1.117888
         timesq |   1.001523    .001342     1.14   0.256      .998896    1.004157
        timecub |   .9999924   6.69e-06    -1.13   0.258     .9999793    1.000006
          _cons |   .0003871   .0006824    -4.46   0.000     .0000122    .0122563
----------------+----------------------------------------------------------------
       /lnsig2u |  -2.025709   1.873117                     -5.696955    1.645537
----------------+----------------------------------------------------------------
        sigma_u |   .3631807   .3401399                      .0579324    2.276794
            rho |   .0742333   .1287256                      .0020362    .7591155
---------------------------------------------------------------------------------

. *       RESULTS:  lndollexchratel--positive and statistically insignificant, did not change significance or sign
> s of other variables
. *                 no bivariate relationship
. 
. * Age of incumbent leader--ageincleader
. * On 10 imputed datasets, testing Model 4 from Table 3.1
. clear

. use revspredictbycntryyr, clear

. drop if indstate==0
(6,647 observations deleted)

. misstable sum ageincleader, gen(miss)
                                                               Obs<.
                                                +------------------------------
               |                                | Unique
      Variable |     Obs=.     Obs>.     Obs<.  | values        Min         Max
  -------------+--------------------------------+------------------------------
  ageincleader |        91              11,651  |     82         11          92
  -----------------------------------------------------------------------------

. mi set wide

. mi register regular lnpopl postcoldwar cowcode year urbancivicny 

. mi register imputed polityl gdppcthl  yrsincleaderinpower v2x_execorr lnoill ageincleader

. mi xtset cowcode year
       panel variable:  cowcode (unbalanced)
        time variable:  year, 1900 to 2014, but with gaps
                delta:  1 unit

. mi impute chained (pmm, knn(10)) polityl gdppcthl yrsincleaderinpower v2x_execorr lnoill ageincleader = lnpopl p
> ostcoldwar, add(10) rseed(1234) force dots chaindots

Conditional models:
    yrsincleader~r: pmm yrsincleaderinpower ageincleader v2x_execorr lnoill polityl gdppcthl lnpopl postcoldwar
                     , knn(10)
      ageincleader: pmm ageincleader yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl lnpopl postcoldwar
                     , knn(10)
       v2x_execorr: pmm v2x_execorr yrsincleaderinpower ageincleader lnoill polityl gdppcthl lnpopl postcoldwar
                     , knn(10)
            lnoill: pmm lnoill yrsincleaderinpower ageincleader v2x_execorr polityl gdppcthl lnpopl postcoldwar
                     , knn(10)
           polityl: pmm polityl yrsincleaderinpower ageincleader v2x_execorr lnoill gdppcthl lnpopl postcoldwar
                     , knn(10)
          gdppcthl: pmm gdppcthl yrsincleaderinpower ageincleader v2x_execorr lnoill polityl lnpopl postcoldwar
                     , knn(10)

Performing chained iterations:
  imputing m=1: burn-in 10 .......... done
  imputing m=2: burn-in 10 .......... done
  imputing m=3: burn-in 10 .......... done
  imputing m=4: burn-in 10 .......... done
  imputing m=5: burn-in 10 .......... done
  imputing m=6: burn-in 10 .......... done
  imputing m=7: burn-in 10 .......... done
  imputing m=8: burn-in 10 .......... done
  imputing m=9: burn-in 10 .......... done
  imputing m=10: burn-in 10 .......... done

Multivariate imputation                     Imputations =       10
Chained equations                                 added =       10
Imputed: m=1 through m=10                       updated =        0

Initialization: monotone                     Iterations =      100
                                                burn-in =       10

           polityl: predictive mean matching
          gdppcthl: predictive mean matching
    yrsincleader~r: predictive mean matching
       v2x_execorr: predictive mean matching
            lnoill: predictive mean matching
      ageincleader: predictive mean matching

------------------------------------------------------------------
                   |               Observations per m             
                   |----------------------------------------------
          Variable |   Complete   Incomplete   Imputed |     Total
-------------------+-----------------------------------+----------
           polityl |      11328          414       414 |     11742
          gdppcthl |      11042          700       700 |     11742
    yrsincleader~r |      11661           81        81 |     11742
       v2x_execorr |      11580          162       162 |     11742
            lnoill |      11560          182       182 |     11742
      ageincleader |      11651           91        91 |     11742
------------------------------------------------------------------
(complete + incomplete = total; imputed is the minimum across m
 of the number of filled-in observations.)

. qui mi xeq 1: twoway (kdensity ageincleader if missageincleader==0) || (kdensity ageincleader if missageincleade
> r==1) || (kdensity ageincleader), legend(label(1 "Observed") label(2 "Imputed") label(3 "Completed"))

. graph export Robustnesstestfiles\Logfiles\mod4impobsageincleader.pdf, replace
(file Robustnesstestfiles\Logfiles\mod4impobsageincleader.pdf written in PDF format)

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny lnpopl c.gdppcthl##c.gdppcthl
>  c.polityl##c.polityl yrsincleaderinpower v2x_execorr lnoill postcoldwar ageincleader, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0028
                                                Largest FMI       =     0.0179
DF adjustment:   Large sample                   DF:     min       =  28,165.75
                                                        avg       =   1.92e+18
                                                        max       =   2.31e+19
Model F test:       Equal FMI                   F(  10, 7.7e+06)  =      11.56
Within VCE type:       Robust                   Prob > F          =     0.0000

                                      (Within VCE adjusted for 164 clusters in cowcode)
---------------------------------------------------------------------------------------
         urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------------+----------------------------------------------------------------
               lnpopl |    1.47734   .1421702     4.06   0.000     1.223393       1.784
             gdppcthl |   1.825404    .241686     4.55   0.000     1.408182    2.366243
                      |
c.gdppcthl#c.gdppcthl |   .9669927   .0089796    -3.61   0.000     .9495522    .9847536
                      |
              polityl |   .9242751   .0303123    -2.40   0.016     .8667333     .985637
                      |
  c.polityl#c.polityl |   .9844544   .0055297    -2.79   0.005     .9736754    .9953527
                      |
  yrsincleaderinpower |   1.026093   .0129234     2.05   0.041     1.001074    1.051738
          v2x_execorr |     5.1004   4.258851     1.95   0.051     .9927686    26.20356
               lnoill |   .8771625   .0259118    -4.44   0.000     .8278185    .9294477
          postcoldwar |   7.490844   3.208853     4.70   0.000     3.235202    17.34443
         ageincleader |   1.017892    .013133     1.37   0.169     .9924745     1.04396
                _cons |   2.47e-06   3.05e-06   -10.46   0.000     2.20e-07    .0000278
----------------------+----------------------------------------------------------------
             /lnsig2u |   -10.7042   21237.29                     -41635.02    41613.61
----------------------+----------------------------------------------------------------
              sigma_u |   .0047382   50.31321                             0           .
                  rho |   .0000136   .2898443                             0           .
---------------------------------------------------------------------------------------

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny ageincleader time1 timesq tim
> ecub, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0000
                                                Largest FMI       =     0.0000
DF adjustment:   Large sample                   DF:     min       =   1.41e+10
                                                        avg       =   5.05e+11
                                                        max       =   1.11e+12
Model F test:       Equal FMI                   F(   4, 2.9e+11)  =       9.29
Within VCE type:       Robust                   Prob > F          =     0.0000

                             (Within VCE adjusted for 164 clusters in cowcode)
------------------------------------------------------------------------------
urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
ageincleader |   1.040042   .0135113     3.02   0.003     1.013895    1.066864
       time1 |   .9491294   .0745398    -0.66   0.506     .8137226    1.107068
      timesq |   1.001571   .0013039     1.21   0.228     .9990186     1.00413
     timecub |   .9999922   6.53e-06    -1.19   0.233     .9999794    1.000005
       _cons |   .0000483   .0000905    -5.30   0.000     1.23e-06    .0018994
-------------+----------------------------------------------------------------
    /lnsig2u |  -1.307099    .905552                     -3.081949    .4677501
-------------+----------------------------------------------------------------
     sigma_u |    .520196   .2355323                      .2141723    1.263487
         rho |   .1412678   .1098536                       .027129     .492513
------------------------------------------------------------------------------

. *       RESULTS:  ageincleader--positive but statistically insignificant; did not change significance or signs o
> f other variables; 
. *                 bivariate specification--positive and significant at the .001 level
. 
. * Logged military expenditures per soldier--lnmilexppersoldthl
. * On 10 imputed datasets, testing Model 4 from Table 3.1
. clear

. use revspredictbycntryyr, clear

. drop if indstate==0
(6,647 observations deleted)

. misstable sum lnmilexppersoldthl, gen(miss)
                                                               Obs<.
                                                +------------------------------
               |                                | Unique
      Variable |     Obs=.     Obs>.     Obs<.  | values        Min         Max
  -------------+--------------------------------+------------------------------
  lnmilexppe~l |     1,817               9,925  |   >500          0    7.790558
  -----------------------------------------------------------------------------

. mi set wide

. mi register regular lnpopl postcoldwar cowcode year urbancivicny 

. mi register imputed polityl gdppcthl  yrsincleaderinpower v2x_execorr lnoill lnmilexppersoldthl

. mi xtset cowcode year
       panel variable:  cowcode (unbalanced)
        time variable:  year, 1900 to 2014, but with gaps
                delta:  1 unit

. mi impute chained (pmm, knn(10)) polityl gdppcthl yrsincleaderinpower v2x_execorr lnoill lnmilexppersoldthl = ln
> popl postcoldwar, add(10) rseed(1234) force dots chaindots

Conditional models:
    yrsincleader~r: pmm yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl lnmilexppersoldthl lnpopl
                     postcoldwar , knn(10)
       v2x_execorr: pmm v2x_execorr yrsincleaderinpower lnoill polityl gdppcthl lnmilexppersoldthl lnpopl
                     postcoldwar , knn(10)
            lnoill: pmm lnoill yrsincleaderinpower v2x_execorr polityl gdppcthl lnmilexppersoldthl lnpopl
                     postcoldwar , knn(10)
           polityl: pmm polityl yrsincleaderinpower v2x_execorr lnoill gdppcthl lnmilexppersoldthl lnpopl
                     postcoldwar , knn(10)
          gdppcthl: pmm gdppcthl yrsincleaderinpower v2x_execorr lnoill polityl lnmilexppersoldthl lnpopl
                     postcoldwar , knn(10)
    lnmilexppers~l: pmm lnmilexppersoldthl yrsincleaderinpower v2x_execorr lnoill polityl gdppcthl lnpopl
                     postcoldwar , knn(10)

Performing chained iterations:
  imputing m=1: burn-in 10 .......... done
  imputing m=2: burn-in 10 .......... done
  imputing m=3: burn-in 10 .......... done
  imputing m=4: burn-in 10 .......... done
  imputing m=5: burn-in 10 .......... done
  imputing m=6: burn-in 10 .......... done
  imputing m=7: burn-in 10 .......... done
  imputing m=8: burn-in 10 .......... done
  imputing m=9: burn-in 10 .......... done
  imputing m=10: burn-in 10 .......... done

Multivariate imputation                     Imputations =       10
Chained equations                                 added =       10
Imputed: m=1 through m=10                       updated =        0

Initialization: monotone                     Iterations =      100
                                                burn-in =       10

           polityl: predictive mean matching
          gdppcthl: predictive mean matching
    yrsincleader~r: predictive mean matching
       v2x_execorr: predictive mean matching
            lnoill: predictive mean matching
    lnmilexppers~l: predictive mean matching

------------------------------------------------------------------
                   |               Observations per m             
                   |----------------------------------------------
          Variable |   Complete   Incomplete   Imputed |     Total
-------------------+-----------------------------------+----------
           polityl |      11328          414       414 |     11742
          gdppcthl |      11042          700       700 |     11742
    yrsincleader~r |      11661           81        81 |     11742
       v2x_execorr |      11580          162       162 |     11742
            lnoill |      11560          182       182 |     11742
    lnmilexppers~l |       9925         1817      1817 |     11742
------------------------------------------------------------------
(complete + incomplete = total; imputed is the minimum across m
 of the number of filled-in observations.)

. qui mi xeq 1: twoway (kdensity lnmilexppersoldthl if misslnmilexppersoldthl==0) || (kdensity lnmilexppersoldthl 
> if misslnmilexppersoldthl==1) || (kdensity lnmilexppersoldthl), legend(label(1 "Observed") label(2 "Imputed") la
> bel(3 "Completed"))

. graph export Robustnesstestfiles\Logfiles\mod4impobslnmilexppersoldthl.pdf, replace
(file Robustnesstestfiles\Logfiles\mod4impobslnmilexppersoldthl.pdf written in PDF format)

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny lnpopl c.gdppcthl##c.gdppcthl
>  c.polityl##c.polityl yrsincleaderinpower v2x_execorr lnoill postcoldwar lnmilexppersoldthl, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0061
                                                Largest FMI       =     0.0302
DF adjustment:   Large sample                   DF:     min       =  10,000.42
                                                        avg       =   4.38e+18
                                                        max       =   5.25e+19
Model F test:       Equal FMI                   F(  10, 1.7e+06)  =      10.49
Within VCE type:       Robust                   Prob > F          =     0.0000

                                      (Within VCE adjusted for 164 clusters in cowcode)
---------------------------------------------------------------------------------------
         urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
----------------------+----------------------------------------------------------------
               lnpopl |   1.527361   .1390771     4.65   0.000     1.277714    1.825785
             gdppcthl |   1.862513   .2379256     4.87   0.000     1.449983     2.39241
                      |
c.gdppcthl#c.gdppcthl |   .9664673   .0089049    -3.70   0.000     .9491706    .9840791
                      |
              polityl |   .9267306   .0300869    -2.34   0.019     .8695983    .9876164
                      |
  c.polityl#c.polityl |     .98539   .0055661    -2.61   0.009     .9745406    .9963602
                      |
  yrsincleaderinpower |   1.035012   .0121603     2.93   0.003     1.011451    1.059123
          v2x_execorr |   5.203127   4.353929     1.97   0.049     1.009222    26.82515
               lnoill |   .8799438   .0265822    -4.23   0.000     .8293559    .9336174
          postcoldwar |   8.102034   3.983082     4.26   0.000     3.091221    21.23528
   lnmilexppersoldthl |   .9323281   .1645292    -0.40   0.691     .6596869    1.317649
                _cons |   4.57e-06   5.88e-06    -9.57   0.000     3.69e-07    .0000567
----------------------+----------------------------------------------------------------
             /lnsig2u |  -10.52715   14639.81                     -28704.02    28682.97
----------------------+----------------------------------------------------------------
              sigma_u |   .0051768   37.89345                             0           .
                  rho |   .0000163    .238501                             0           .
---------------------------------------------------------------------------------------

. mi estimate, ni(10) post dots eform saving(miest, replace): xtcloglog urbancivicny lnmilexppersoldthl time1 time
> sq timecub, vce(robust)

Imputations (10):
  .........10 done

Multiple-imputation estimates                   Imputations       =         10
Random-effects complementary log-log model      Number of obs     =     11,742

Group variable: cowcode                         Number of groups  =        164
Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         10
Integration points = 12                                       avg =       71.6
                                                              max =        115
                                                Average RVI       =     0.0021
                                                Largest FMI       =     0.0117
DF adjustment:   Large sample                   DF:     min       =  65,920.03
                                                        avg       =   2.30e+10
                                                        max       =   7.01e+10
Model F test:       Equal FMI                   F(   4, 3.1e+06)  =       8.08
Within VCE type:       Robust                   Prob > F          =     0.0000

                                   (Within VCE adjusted for 164 clusters in cowcode)
------------------------------------------------------------------------------------
      urbancivicny |     exp(b)   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------------+----------------------------------------------------------------
lnmilexppersoldthl |   .7722179   .0854112    -2.34   0.019     .6217144    .9591549
             time1 |   .9470921   .0756084    -0.68   0.496     .8099141    1.107504
            timesq |   1.001701   .0013369     1.27   0.203     .9990838    1.004324
           timecub |   .9999918   6.72e-06    -1.22   0.223     .9999786    1.000005
             _cons |   .0005209    .000901    -4.37   0.000     .0000176    .0154501
-------------------+----------------------------------------------------------------
          /lnsig2u |  -2.280898   2.329328                     -6.846296    2.284501
-------------------+----------------------------------------------------------------
           sigma_u |   .3196755   .3723145                      .0326096    3.133813
               rho |   .0584917   .1282771                       .000646    .8565346
------------------------------------------------------------------------------------

. *       RESULTS:  lnmilexppersoldthl--negative and statistically insignificant, did not change significance or s
> igns of other variables
. *                 bivariate relationship is negative and statistically significant at the .01 level
. 
. * ++++++++++++++
. * Further tests
. * ++++++++++++++
. * Country fixed effects are inappropriate given the rare event character of the data
. * Introduction of dummy controls for world regions (North America and Western Europe as omitted regions)
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar africa latam mena formercomm soseasia eastasia if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,516
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       67.0
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(15)     =     102.07
Log pseudolikelihood  = -261.16623              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 157 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.462708   .1878752     2.96   0.003     1.137173    1.881433
           gdppcthl |   1.629852   .2453428     3.25   0.001     1.213435    2.189174
          gdppcthl2 |   .9730001   .0093079    -2.86   0.004     .9549268    .9914154
            polityl |   .9287636   .0322817    -2.13   0.033     .8675996    .9942396
           polityl2 |   .9839627    .005875    -2.71   0.007      .972515    .9955451
yrsincleaderinpower |   1.037698   .0130517     2.94   0.003      1.01243    1.063597
        v2x_execorr |   3.849391   3.294112     1.58   0.115     .7194113    20.59714
             lnoill |   .8870057   .0283996    -3.74   0.000     .8330541    .9444514
        postcoldwar |   7.027839   3.363807     4.07   0.000     2.750442    17.95731
             africa |   .7249007   .5848672    -0.40   0.690     .1491107      3.5241
              latam |   .8920197    .794307    -0.13   0.898     .1557451    5.108983
               mena |    .982283   .8461169    -0.02   0.983     .1815643    5.314261
         formercomm |   1.792928   1.453401     0.72   0.471     .3660612    8.781567
           soseasia |   1.611877   1.263028     0.61   0.542     .3470168    7.487092
           eastasia |   .8627882   .8387889    -0.15   0.879     .1283452     5.80001
              _cons |   .0000108   .0000192    -6.45   0.000     3.35e-07    .0003491
--------------------+----------------------------------------------------------------
           /lnsig2u |  -11.35729   48079.98                     -94246.39    94223.67
--------------------+----------------------------------------------------------------
            sigma_u |   .0034182   82.17313                             0           .
                rho |   7.10e-06   .3415078                             0           .
-------------------------------------------------------------------------------------

. cloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldw
> ar africa latam mena formercomm soseasia eastasia if indstate==1, vce(robust) eform nolog

Complementary log-log regression                Number of obs     =     10,516
                                                Zero outcomes     =     10,466
                                                Nonzero outcomes  =         50

                                                Wald chi2(15)     =      79.75
Log pseudolikelihood = -261.16621               Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.462713   .2006565     2.77   0.006     1.117868    1.913938
           gdppcthl |   1.629851   .2689151     2.96   0.003     1.179519    2.252116
          gdppcthl2 |   .9730002   .0101642    -2.62   0.009     .9532814     .993127
            polityl |   .9287656   .0325009    -2.11   0.035     .8672003    .9947016
           polityl2 |   .9839629   .0055251    -2.88   0.004     .9731934    .9948517
yrsincleaderinpower |   1.037697   .0125916     3.05   0.002     1.013309    1.062672
        v2x_execorr |   3.849529   3.102885     1.67   0.094     .7930538    18.68583
             lnoill |   .8870052   .0326244    -3.26   0.001      .825313    .9533089
        postcoldwar |   7.027597   2.934559     4.67   0.000     3.100025     15.9312
             africa |   .7249261   .5381194    -0.43   0.665      .169215    3.105622
              latam |   .8920288   .7587564    -0.13   0.893     .1684031    4.725065
               mena |   .9823235   .8173828    -0.02   0.983      .192299    5.018015
         formercomm |   1.792948   1.449837     0.72   0.470     .3675006    8.747367
           soseasia |   1.611927   1.169891     0.66   0.511     .3886577    6.685339
           eastasia |   .8628101   .8109212    -0.16   0.875     .1367428    5.444099
              _cons |   .0000108   .0000201    -6.15   0.000     2.83e-07    .0004138
-------------------------------------------------------------------------------------

. firthlogit urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postco
> ldwar africa latam mena formercomm soseasia eastasia if indstate==1, or

initial:       penalized log likelihood = -282.06973
rescale:       penalized log likelihood = -282.06973
Iteration 0:   penalized log likelihood = -282.06973  (not concave)
Iteration 1:   penalized log likelihood = -253.23825  
Iteration 2:   penalized log likelihood = -233.53303  (not concave)
Iteration 3:   penalized log likelihood = -229.96972  
Iteration 4:   penalized log likelihood = -227.52923  
Iteration 5:   penalized log likelihood = -227.42131  
Iteration 6:   penalized log likelihood = -227.42053  
Iteration 7:   penalized log likelihood = -227.42053  

                                                Number of obs     =     10,516
                                                Wald chi2(15)     =      87.22
Penalized log likelihood = -227.42053           Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
       urbancivicny | Odds Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.453318   .2002074     2.71   0.007      1.10943    1.903799
           gdppcthl |   1.535879   .2684037     2.46   0.014     1.090447    2.163263
          gdppcthl2 |   .9783472   .0121602    -1.76   0.078     .9548016    1.002473
            polityl |   .9311816   .0281669    -2.36   0.018     .8775801     .988057
           polityl2 |    .984334   .0056731    -2.74   0.006     .9732776    .9955161
yrsincleaderinpower |   1.038965   .0146666     2.71   0.007     1.010613    1.068113
        v2x_execorr |   3.591411   2.452695     1.87   0.061     .9417835    13.69554
             lnoill |   .8898207   .0355955    -2.92   0.004     .8227197    .9623945
        postcoldwar |   6.743657   3.036731     4.24   0.000     2.789918    16.30044
             africa |   .5871096   .4765144    -0.66   0.512     .1196356     2.88123
              latam |   .7928475   .6393778    -0.29   0.773     .1632124    3.851466
               mena |   .8405477   .6789286    -0.22   0.830     .1725951    4.093514
         formercomm |   1.498096   1.127824     0.54   0.591     .3425487    6.551743
           soseasia |   1.360755   1.054702     0.40   0.691      .297869    6.216337
           eastasia |   .8377227   .7539963    -0.20   0.844     .1435398    4.889093
              _cons |   .0000181   .0000296    -6.68   0.000     7.35e-07    .0004451
-------------------------------------------------------------------------------------

. * Also rotated each region into the regression separately
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar africa if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,516
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       67.0
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      99.00
Log pseudolikelihood  = -262.96867              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 157 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.471996    .146196     3.89   0.000     1.211621    1.788325
           gdppcthl |   1.665063   .2251121     3.77   0.000     1.277469    2.170256
          gdppcthl2 |   .9726031    .008213    -3.29   0.001     .9566385    .9888342
            polityl |   .9275819   .0309961    -2.25   0.024     .8687774    .9903667
           polityl2 |   .9834074   .0055368    -2.97   0.003     .9726151    .9943195
yrsincleaderinpower |   1.033904   .0123916     2.78   0.005       1.0099    1.058479
        v2x_execorr |   4.756877    4.15855     1.78   0.074     .8574046    26.39113
             lnoill |   .8802943   .0259619    -4.32   0.000     .8308527     .932678
        postcoldwar |   7.630041   3.530168     4.39   0.000     3.081088    18.89512
             africa |   .5880206   .2766986    -1.13   0.259     .2338038    1.478881
              _cons |   .0000106   .0000151    -8.04   0.000     6.48e-07    .0001723
--------------------+----------------------------------------------------------------
           /lnsig2u |  -9.935938   8068.035                     -15822.99    15803.12
--------------------+----------------------------------------------------------------
            sigma_u |   .0069573   28.06572                             0           .
                rho |   .0000294   .2373945                             0           .
-------------------------------------------------------------------------------------

. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar latam if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,516
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       67.0
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      96.23
Log pseudolikelihood  = -263.39649              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 157 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.501281   .1426572     4.28   0.000     1.246171    1.808615
           gdppcthl |   1.841609   .2378138     4.73   0.000     1.429811    2.372009
          gdppcthl2 |   .9665163   .0090336    -3.64   0.000     .9489719    .9843849
            polityl |   .9306346   .0322255    -2.08   0.038     .8695695    .9959879
           polityl2 |   .9842977   .0055195    -2.82   0.005      .973539    .9951753
yrsincleaderinpower |    1.03248   .0126486     2.61   0.009     1.007985    1.057571
        v2x_execorr |    5.23742   4.492147     1.93   0.054     .9750794    28.13163
             lnoill |   .8800973   .0253804    -4.43   0.000     .8317324    .9312747
        postcoldwar |   6.562165   2.929475     4.21   0.000     2.735613    15.74126
              latam |   .7214008   .3881917    -0.61   0.544     .2512685    2.071167
              _cons |   6.77e-06   9.13e-06    -8.83   0.000     4.83e-07    .0000951
--------------------+----------------------------------------------------------------
           /lnsig2u |  -11.03084   26868.29                     -52671.92    52649.86
--------------------+----------------------------------------------------------------
            sigma_u |   .0040242   54.06222                             0           .
                rho |   9.84e-06   .2645152                             0           .
-------------------------------------------------------------------------------------

. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar mena if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,516
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       67.0
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =     100.78
Log pseudolikelihood  = -263.49719              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 157 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.502875   .1507499     4.06   0.000     1.234641    1.829384
           gdppcthl |   1.840792   .2321322     4.84   0.000     1.437687    2.356921
          gdppcthl2 |   .9666116   .0086621    -3.79   0.000     .9497824     .983739
            polityl |   .9251321    .030614    -2.35   0.019     .8670341    .9871231
           polityl2 |   .9840572   .0054683    -2.89   0.004     .9733978    .9948334
yrsincleaderinpower |   1.033737   .0128244     2.67   0.007     1.008904     1.05918
        v2x_execorr |   4.907596   4.149956     1.88   0.060     .9355707    25.74311
             lnoill |   .8823319    .027665    -3.99   0.000      .829742    .9382549
        postcoldwar |   7.061635   3.121188     4.42   0.000     2.969482    16.79306
               mena |     .80638   .3189978    -0.54   0.586     .3713734    1.750929
              _cons |   6.38e-06   8.69e-06    -8.78   0.000     4.42e-07    .0000921
--------------------+----------------------------------------------------------------
           /lnsig2u |  -9.944091   8416.862                     -16506.69     16486.8
--------------------+----------------------------------------------------------------
            sigma_u |    .006929   29.16005                             0           .
                rho |   .0000292   .2456475                             0           .
-------------------------------------------------------------------------------------

. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar formercomm if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,516
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       67.0
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      99.69
Log pseudolikelihood  =  -262.4573              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 157 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.544272   .1546664     4.34   0.000      1.26903    1.879212
           gdppcthl |   1.704235    .259206     3.51   0.000     1.264928    2.296113
          gdppcthl2 |   .9702702   .0095211    -3.08   0.002     .9517875    .9891118
            polityl |   .9301974   .0310933    -2.16   0.030     .8712091    .9931798
           polityl2 |   .9842548   .0055451    -2.82   0.005     .9734465    .9951832
yrsincleaderinpower |   1.035382   .0127544     2.82   0.005     1.010683    1.060684
        v2x_execorr |   4.581523   3.894298     1.79   0.073     .8659402    24.23995
             lnoill |   .8817939   .0268597    -4.13   0.000     .8306905    .9360413
        postcoldwar |    6.88142   3.137401     4.23   0.000     2.815758    16.81748
         formercomm |   1.756216   .7030973     1.41   0.160     .8013113     3.84906
              _cons |   5.42e-06   7.37e-06    -8.91   0.000     3.77e-07     .000078
--------------------+----------------------------------------------------------------
           /lnsig2u |  -10.20438   11742.96                     -23025.98    23005.57
--------------------+----------------------------------------------------------------
            sigma_u |   .0060834   35.71854                             0           .
                rho |   .0000225   .2641811                             0           .
-------------------------------------------------------------------------------------

. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar soseasia if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,516
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       67.0
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      97.20
Log pseudolikelihood  = -262.97909              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 157 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.459863     .14118     3.91   0.000     1.207799    1.764534
           gdppcthl |    1.82911   .2332419     4.74   0.000     1.424615    2.348455
          gdppcthl2 |    .967506   .0086934    -3.68   0.000     .9506164    .9846957
            polityl |    .924349   .0308894    -2.35   0.019      .865747    .9869177
           polityl2 |   .9841329    .005409    -2.91   0.004     .9735883    .9947916
yrsincleaderinpower |   1.034239   .0128491     2.71   0.007      1.00936    1.059732
        v2x_execorr |   4.487981   3.859622     1.75   0.081     .8318125    24.21456
             lnoill |   .8819112    .026826    -4.13   0.000     .8308698    .9360881
        postcoldwar |   6.767749   2.988205     4.33   0.000     2.848447    16.07979
           soseasia |   1.584282   .5980198     1.22   0.223     .7560129    3.319984
              _cons |   8.24e-06    .000011    -8.81   0.000     6.10e-07    .0001114
--------------------+----------------------------------------------------------------
           /lnsig2u |  -10.84116   22089.28                     -43305.04    43283.36
--------------------+----------------------------------------------------------------
            sigma_u |   .0044246   48.86798                             0           .
                rho |   .0000119   .2628865                             0           .
-------------------------------------------------------------------------------------

. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar eastasia if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,516
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       67.0
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(10)     =      96.87
Log pseudolikelihood  = -263.31473              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 157 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.585069   .1863705     3.92   0.000     1.258824    1.995866
           gdppcthl |   1.841917   .2369554     4.75   0.000     1.431417    2.370139
          gdppcthl2 |   .9671254   .0087874    -3.68   0.000     .9500549    .9845027
            polityl |   .9236055   .0304185    -2.41   0.016     .8658698    .9851911
           polityl2 |   .9845575   .0055164    -2.78   0.005     .9738047    .9954289
yrsincleaderinpower |    1.03368   .0124909     2.74   0.006     1.009486    1.058454
        v2x_execorr |   5.276042   4.554128     1.93   0.054     .9717971    28.64448
             lnoill |   .8709142    .026444    -4.55   0.000      .820597    .9243168
        postcoldwar |   6.796266   2.976404     4.38   0.000     2.880642    16.03435
           eastasia |   .6054925   .3915752    -0.78   0.438     .1704639    2.150727
              _cons |   3.80e-06   5.60e-06    -8.48   0.000     2.12e-07    .0000681
--------------------+----------------------------------------------------------------
           /lnsig2u |  -10.58234   16275.47                     -31909.92    31888.75
--------------------+----------------------------------------------------------------
            sigma_u |   .0050359   40.98046                             0           .
                rho |   .0000154   .2509101                             0           .
-------------------------------------------------------------------------------------

. *  RESULT
. *    --none of the regional dummies are statistically significant
. *    --no sign change or change in statistical significance for any of the variables
. 
. * Alternative sample urbancivicaltny with Model 4
. *   Includes an additional 11 cases that were quasi-revolutionary but urban civic in character
. xtcloglog urbancivicaltny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill post
> coldwar if indstate==1, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,516
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       67.0
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(9)      =     143.87
Log pseudolikelihood  = -343.03504              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 157 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
    urbancivicaltny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.445421   .1404029     3.79   0.000     1.194846    1.748546
           gdppcthl |   1.786938   .1815527     5.71   0.000     1.464292    2.180678
          gdppcthl2 |   .9732635   .0072907    -3.62   0.000     .9590782    .9876585
            polityl |   .9155248   .0242855    -3.33   0.001     .8691423    .9643825
           polityl2 |   .9828178   .0044201    -3.85   0.000     .9741926    .9915195
yrsincleaderinpower |   1.032059   .0106202     3.07   0.002     1.011452    1.053085
        v2x_execorr |   3.154828   1.705282     2.13   0.034     1.093641    9.100735
             lnoill |   .8660159   .0232491    -5.36   0.000     .8216266    .9128034
        postcoldwar |   3.632201   1.230706     3.81   0.000     1.869629    7.056417
              _cons |   .0000285   .0000319    -9.33   0.000     3.16e-06    .0002563
--------------------+----------------------------------------------------------------
           /lnsig2u |  -11.20119    21260.4                     -41680.82    41658.42
--------------------+----------------------------------------------------------------
            sigma_u |   .0036957   39.28558                             0           .
                rho |   8.30e-06   .1765221                             0           .
-------------------------------------------------------------------------------------

. * failed quadchk--using pooled model
. cloglog urbancivicaltny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postco
> ldwar if indstate==1, vce(robust) eform nolog

Complementary log-log regression                Number of obs     =     10,516
                                                Zero outcomes     =     10,450
                                                Nonzero outcomes  =         66

                                                Wald chi2(9)      =     113.69
Log pseudolikelihood = -343.03499               Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
                    |               Robust
    urbancivicaltny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.445414   .1386663     3.84   0.000     1.197655    1.744426
           gdppcthl |   1.786919   .1986485     5.22   0.000     1.437072    2.221935
          gdppcthl2 |   .9732644   .0075477    -3.49   0.000     .9585831    .9881705
            polityl |   .9155183   .0254205    -3.18   0.001     .8670265    .9667223
           polityl2 |   .9828182   .0043862    -3.88   0.000     .9742588    .9914527
yrsincleaderinpower |   1.032062   .0103677     3.14   0.002     1.011941    1.052584
        v2x_execorr |   3.154979   1.877951     1.93   0.054     .9824994    10.13119
             lnoill |   .8660128   .0253133    -4.92   0.000     .8177939    .9170747
        postcoldwar |   3.632401   1.067904     4.39   0.000     2.041483    6.463115
              _cons |   .0000284   .0000318    -9.36   0.000     3.18e-06    .0002545
-------------------------------------------------------------------------------------

. *  RESULT
. *    --no sign change or change in statistical significance for all variables
. 
. * Excluded Polity=0 in Model 4
. xtcloglog urbancivicny lnpopl gdppcthl gdppcthl2 polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcol
> dwar if indstate==1 & polityl~=0, vce(robust) eform nolog

Calculating robust standard errors:

Random-effects complementary log-log model      Number of obs     =     10,390
Group variable: cowcode                         Number of groups  =        157

Random effects u_i ~ Gaussian                   Obs per group:
                                                              min =         21
                                                              avg =       66.2
                                                              max =        114

Integration method: mvaghermite                 Integration pts.  =         12

                                                Wald chi2(9)      =      91.25
Log pseudolikelihood  = -255.37457              Prob > chi2       =     0.0000

                                     (Std. Err. adjusted for 157 clusters in cowcode)
-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.493234   .1465466     4.09   0.000     1.231943    1.809944
           gdppcthl |   1.832062   .2502782     4.43   0.000     1.401707    2.394544
          gdppcthl2 |   .9670077   .0092779    -3.50   0.000     .9489933    .9853641
            polityl |   .9278296   .0308007    -2.26   0.024     .8693833     .990205
           polityl2 |   .9858537   .0055537    -2.53   0.011     .9750285    .9967991
yrsincleaderinpower |   1.030799   .0121476     2.57   0.010     1.007263    1.054885
        v2x_execorr |   6.863958   5.988571     2.21   0.027     1.241452    37.95065
             lnoill |   .8762981   .0270052    -4.28   0.000     .8249356    .9308586
        postcoldwar |   6.505677   2.855933     4.27   0.000     2.751837    15.38021
              _cons |   5.49e-06   7.68e-06    -8.66   0.000     3.55e-07    .0000851
--------------------+----------------------------------------------------------------
           /lnsig2u |  -8.586156   1807.283                     -3550.795    3533.623
--------------------+----------------------------------------------------------------
            sigma_u |   .0136628   12.34627                             0           .
                rho |   .0001135   .2050495                             0           .
-------------------------------------------------------------------------------------

. *  RESULT
. *    --no sign change or change in statistical significance for all variables
. 
. * ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
. * Substitutability of percurbanl, under5mortl, and youthpercl for gdppcthl 
. *       without affecting other variables
. * ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
. cloglog urbancivicny lnpopl percurbanl polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldwar if in
> dstate==1, vce(robust) eform nolog

Complementary log-log regression                Number of obs     =     10,887
                                                Zero outcomes     =     10,837
                                                Nonzero outcomes  =         50

                                                Wald chi2(8)      =      62.59
Log pseudolikelihood = -271.18246               Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.601348    .180299     4.18   0.000     1.284243    1.996752
         percurbanl |    1.02456   .0096048     2.59   0.010     1.005906    1.043558
            polityl |   .9232448   .0315287    -2.34   0.019     .8634724     .987155
           polityl2 |   .9846334   .0049246    -3.10   0.002     .9750286    .9943329
yrsincleaderinpower |   1.027622    .012095     2.31   0.021     1.004187    1.051603
        v2x_execorr |   4.253801   3.254018     1.89   0.058     .9498083    19.05103
             lnoill |   .9080004   .0299803    -2.92   0.003      .851101    .9687037
        postcoldwar |   6.363599   2.734185     4.31   0.000     2.741419    14.77169
              _cons |   4.87e-06   7.54e-06    -7.89   0.000     2.33e-07    .0001016
-------------------------------------------------------------------------------------

. cloglog urbancivicny lnpopl under5mortl polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldwar if i
> ndstate==1, vce(robust) eform nolog

Complementary log-log regression                Number of obs     =     10,944
                                                Zero outcomes     =     10,896
                                                Nonzero outcomes  =         48

                                                Wald chi2(8)      =      61.67
Log pseudolikelihood = -258.69193               Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.428122   .1641234     3.10   0.002     1.140101    1.788906
        under5mortl |    .990033   .0029765    -3.33   0.001     .9842164     .995884
            polityl |   .9296004   .0317962    -2.13   0.033      .869324    .9940562
           polityl2 |   .9833038   .0052843    -3.13   0.002     .9730012    .9937156
yrsincleaderinpower |   1.025574   .0137204     1.89   0.059     .9990323    1.052822
        v2x_execorr |   5.424964   4.186812     2.19   0.028     1.195273    24.62219
             lnoill |   .9270108   .0291581    -2.41   0.016     .8715879    .9859579
        postcoldwar |   3.074548   1.403362     2.46   0.014     1.256767    7.521559
              _cons |   .0001605   .0002317    -6.05   0.000     9.49e-06    .0027158
-------------------------------------------------------------------------------------

. cloglog urbancivicny lnpopl youthpercl polityl polityl2 yrsincleaderinpower v2x_execorr lnoill postcoldwar if in
> dstate==1, vce(robust) eform nolog

Complementary log-log regression                Number of obs     =      8,131
                                                Zero outcomes     =      8,084
                                                Nonzero outcomes  =         47

                                                Wald chi2(8)      =      68.46
Log pseudolikelihood =  -246.7512               Prob > chi2       =     0.0000

-------------------------------------------------------------------------------------
                    |               Robust
       urbancivicny |     exp(b)   Std. Err.      z    P>|z|     [95% Conf. Interval]
--------------------+----------------------------------------------------------------
             lnpopl |   1.473864   .1709468     3.34   0.001     1.174169    1.850054
         youthpercl |    .907659   .0240812    -3.65   0.000     .8616669     .956106
            polityl |   .9241333   .0315377    -2.31   0.021     .8643423    .9880602
           polityl2 |   .9776625   .0058983    -3.74   0.000     .9661702    .9892915
yrsincleaderinpower |   1.033716   .0132226     2.59   0.010     1.008122    1.059959
        v2x_execorr |   6.758419   5.816343     2.22   0.026     1.251109    36.50859
             lnoill |   .9263283   .0310593    -2.28   0.022     .8674104    .9892482
        postcoldwar |   4.631505   2.078196     3.42   0.001     1.922121    11.15998
              _cons |   .0006631   .0010703    -4.53   0.000      .000028    .0156859
-------------------------------------------------------------------------------------

. 
. log close
      name:  <unnamed>
       log:  C:\Users\mbeissin\Desktop\Stata files for book\Robustnesstestfiles\Logfiles\robustnesstestschapter3.l
> og
  log type:  text
 closed on:  26 Jan 2022, 10:21:02
------------------------------------------------------------------------------------------------------------------
